AND CONSTHUCTILN OF A FIVE INCH CEHTRIFUGALÜÜLECULAR STILL

brJohn B? éihlpp, Jr.

A Thesis Subnitted in Partial Fulfillment of the

Requirements for the Degree of

MASTER OF SCIEJCE

in

CHEEICAL ENGINEERING

Approved:^ 4- - Ü

r — — - 4 - —*

In Charge of Investigation#**7 ;

Hea;1 of r DepartmentA! zxß Ä

eering Division1 Ü _ g ,, Ü

Chalznan, Graduate IiätteeO U[

Virginia Polyteehnic Institute

Blacksburg, Virginial9l•8

TABLE OF CONTEHTSPage

I. IHTRCDUCTION ........................................ 1

II. LITERATUHE REVIEW ................................... 3

Developnent ot Hnlecular Distillation ............ 3

Th•ory*o£ Molecular Distillation ................. 5Still Pertonnance ................................ 8

Types of Stills .................................. 13

Comparison of Stille ............................. 16

Elimination Curve Theory ......................... 18

Auziliary Vaeunn Equipnnt ....................... 29

Future of Molecular üistillation ................. 33

3 111.................................... 35A. Purpose ot Stuy ............................. 35

B. Plan of Investigation ........................ 35

C. äateriale .................................... 36

D. Apparatus .................................... 37

E. Design and Construction ...................... A3General ................................... A3

Frame ..................................... A3

Assmbly .................................. 53

Auxiliary Costruction .................... 60

F. Methods of Procedure ......................... 62

G. Data an Results ............................. 66

IV, DISCUSSION ,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,, 70

A, Discussion of Results and Heconendations ,,, 70

General ,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,, 70

Vibration Eneeuntered ,,,,,,,,,,.,,,,,,,., 70

Recommendatioe ,,,,,,,,,,,,,,,,,,,,,,, 71

Paeking Nut ,,,,,,,,,,,.,,,,,,,,,,,,,,,,,, 72

Recounendatien ,,,,,,,,,,,,,,,,,,,,,,,, 7k

Distillate Drauff ,,,,,,,,,,,,,,,,,,,,,,, 75Vacuun Lines to Distillate and Reeidue

Reeeivers ,,,,,,,,,,,,,,,,,,,,,,,.,,,,, 76

Reeoumendation ,,,,,,,,,,,,,,,..,,,.,,, 77

Recireulstion of Feed ,,,,,,,,,,,,,,,,,,,, 77

Degassing ,,,,,,,,,,,,.,,,,,,,,,,,,,,,,,,, 78

Reeeumendation ,,,.,,.,.,,.,,,,,,,.,,,, 79

Advantagee of a Reeireulation Systen ,,,,, 80

Sanpling ,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,, 81 '

Reeonnendatio ,,,,,,,,,,.,„,,,,,,,,,,, 82

Feed Rate ,,,,,,,,,,,.,,,,,,,,,,,,,,,,,,,, 83

Degassing of the Heater Unit ,,,,,,,,,,,,, 85

Reeommendation ,,,.,,,,,,,,,,,,,,,,,,,, 86

Heat Transfer freu the Heater ,„,,,,,,,,,, 86

Heat Transfer to the Base Plate ,,,,,,,,,, 86

Reeoumendation ,,,,,,,,,,..,,,,,,,,,,,, 86

Effect of Heat on the Sealing Materials ,,, 87

Recoumendation ,,•,,,,,„,,,,,.,•,,,,,,,, 87

Charring of the Feed on the Evaporetor ,,,• 88

Leak Finding ,,,•,„,,,,,,,,,,,,•,,,,,,,,•,, 88

Use of the Tesla Coil ,,•,,,,,,,,,,,,,,,,,, 88

Isolation Method ,,,,,.,,,,,,,„,,,,,,,,,,,, 89

Becoumendation ,,,,,,,,,,,,,.,,,,,,,,,,, 90

Use of Tygon Tuing ,,„,,,,,,,,,,,,,,,,,,., 90

Reeemendation ,,,,,,,,,,,.......,.,.,,.

91EvacuatiouTime ,,,,,,,,,,,,,,,,,,•,,,,,,•, 91

Hecoumendation ,,,,,,.,,,,,,,,.,,,,,.,,• 92

Sealing of Joint: ,,,,.,,,,,,,,,.,,,,,,,.,, 92

Recoumeudation ,,,,,.,,,..,,,,,,,,,,,,,, 93

Diffusio Pep ,,,,,,,,,,,,,,,,,,,,,,•,,,,, 9h

Beeemnendation ,,,,,,,,,,,,,,,,,.,,,„,,, 95

B•11—jar Support .....,,.,,,,,,.,,.....,,,, 96

Reeonnendatione ,,,,.„,,,.....,.,,,,,,,. 96

Distillation ,,,,,,,,.,.,,,,,.,,,,,,,,,•,,. 98

B, Limitations ,,,,,,,,,,,,,,,,,,„,,,,,,,,,,,,,„, 100

V, CONCLUSIONS „,,,,,,,,,.,,,,,,,,,,,,,,.,,,,,,,,,,,,,,, 102

v1. smouuax ............................................. 103VII, BIBLIOGHAPH! ..,,..,.,..,,,,,,,,,,,,,.,,.,..,•,,,,,,, 10h

VIII, ACKNOwLEDGEEHTS ,.,,,,,,,,.,,,,,,,,,,,,,,,,,,,,,,,,, 109

FIGUHESFigure Page

1. Yield vs. Tmperature Chart and Elhmination Curve .... 20

2. Effect of Initial Temperature on Eliminstion Curve: .. 21

3. Effect of Time of Diatillation on EliminationCurve: ............................................ 23

A. Effect of Heat of Veporisation on ElininatienCurves ............................................ 2A

5. Elimination Curve: for Various Conditions ............ 26

6. Elimination Curve: for Continuous DietillabilityChange ............................................ 27'

7. Botors ............................................... AA

8. Collecting Cutter .................................... ÄS g

9. Bearing .............................................. A6

10. Base Plate ........................................... A7

11. Rotor Plate an Shaft ................................ A812. Heater Shell ......................................... A9

13. Frame ................................................ 52

1A. Still Head Assembly .................................. 55

15. Still Assembly ....................................... 51

16. Proposed Sampling System ............................. SA

17. Proposed Bell·jar Support ............................ 97

1. lNTRwuUCTION

Molecular distillation is that process of free transfer of

molecules uner high vacuum from an evaporator to a condenser, where

the distance of travel is within the limits of the mean free path of

the vapor molecules in the residual gas. Under these conditions,

distillations can be made at temperatures 100 to l50°C. lower than

ordinary distillations.

Many of the so-called "udietillables“ can be distilled under

conditions existing in the molecular distillstion process. The nat-T ural fats and waxee, sugar derivatives, petroleum residues, plastics,

and plasticizers now fall within the scope of molecular distillation.

The use of the molecular distillation process is now limited

by the high costs involved. Thermal efficiency is low, only 2Z of the

total input heat is utilized for distillation. The separating pouersof molecular stills are low, never more than that of one ”theoretical

plate“. Only those materials that can bear the high process costs are

now beng produced conaercially, the production of vitamine being the

most prominent.

Considerable work has been done relative to the developnent of

mlecular stille. Advancement to the present day centrifugal stille

has been accomplished relatively recently. Many problems are yet un·

solved. Much work must be done before molecular distillation can be

placed on a level with other methods of distillation. An increase inthermal efficiency, an increase in separating pouers, a decrease in

..2..

film thickness, and a lowering of the costa of the process are nec-

eesary.

The purpose of this investigation was to design and construct

a five inch centrifugal molecular still. Four interehangeable rotors

were wployed, each of different thickness, different diueter, and

having e different angle of inclination. The effect of angle of in-

clination on film thicknees can then be studied. Although no data

hae been published relative to still perfornxance of five inch c¤¤•

nercial stille, data obtained from the operation of the still eon··

strueted in this investigation can be caapared when such data ie,T

published.

.3-

II. REVIEM UF LITERATURE

Devcloggggg of äglggular Dgegillgtgg4

Molecular distillation as it is known today had its beginningin 1922,(5) Bronsted and Hevesy,(5)(16) in etteepting to separate theisotopes of nercury, produced a true nolecular etil1.(A)(lh) I¤V••$iga·tioe previous to 1922 have ben classified ae atteets at moleculardistillation, but had no real success. These include the works ofKrafft,(16)(12) Ch•vruel,(l6)(12) and Caldwell and Hurtley.(16)(1¤)(12)

About the eane time Bronsted and Heveey were doing their work,Waterman and co~uorkcre(16)(37) were investigating the possibility ofhigh.vacuu¤ dietillation. following thie, in the eiddle 1920*s,Burch(6) began some investigations in high vacun distillation, usingan original still, not following conventional designs. He showed thatprevios “udistil1able“ petroleun residues could he distilled,(7)giving a suhstantial portion of distillate. In England the Metropoli-tan-Vickere Electrical Conpany, Ltd., and lnperisl Chemical Industries,Ltd., followed up the work of Burch. Bancroft, Burrowe, and Faucett

E were the chief contributors.(l6)(12) The chief cotribtion of these‘ investigators was a high vacun technology relative to design of appa-

retua, including equations where natural nolecular paths were equatedto design and dimensions of apparatus through the kinetic theory ofgasee.

.g.

In Holland, additional work was being carried on hy Waterman

and co·workers. Development and applicatio of pieces of high vac-

_ uum equipnent was their chief contribution. This was perhaps the

earliest really high vacuun distillation of natural oils that hadx

taken place.(l6)

A little work was done at this time in America hy Wilson(“o)

in petroleun lahoratories. These experieents were along the same

» line as Burch's and were short lived. Carothers and Hill(ll) carried

out experiments relative to ehenical reaction: taking place in high

vacuum, such as polymerization. At the Bureau of Standards, Wash-

burn(33) did considerable work on speed of diatillation and degree of

separation which could be expected in the distillation of binary

mixtures. Hickman at Eastman Kodak Company carried ot some experi-

ments on the distillation of plasticizers for photographie film.

Fre this point, efforts to concentrate Vitamin A played a

big part in the developnent of molecular distillation. Hickman,

working indepedently on the fractionation of cod-liver oil by selec-

ular distillation, showed that a very rich distillate could be ob-

tained. Carr and Jew•ll(28) in Englad made advance: in concentrat·

ing the vitamin in the residue. The iaportance of these experinente

is evident by the interest shown by the Eastman Kodak Company and

General hills, Incorporated, in America, an British drug houses in

England, the fonner establishing a separate comany called Distilla-

tion Products, Incorporated, to make further studies.

-5-

Waterman andFaucett(16)(l2)(15) intensively covered the

production of "stand oils” by the distillation of polyenmized oils

of high iodine value to remove the unpolymerized portion„(l6)(12)(15)

Much of the theory and knowledge used today was presents at

this time by Faw¤ett•

Except for one major change, little has been done since 1939

on the design of stills. In this country, Hiekman, at Distillation

Products, Incorporated, Bochester, N. Y., has developed a centrifugal

still which has greatly increased the scope of moleoular distillation.

At the present time there are three chief geographical areas of n-

deavor, America, England and Holland. In Aurica, Hiekman and wash-

burn are the chief contributors; in England, Burch and Faucettg ad

in Holland, Waterman•’ V

fTheogg of Molecular Distillation

Definition: Molecular distillation is defined by Jeans(27)

1 as that process of free transfer of molecules under high Vacuum

from evaporator to condenser where the distance of transfer is

within the limits of the mean free path of the vapor molecules in

the residual gas. Uder such conitions the distillation tempera-

ture of high molecular weight compouds can be lowered 100 to

l50°G.

Hate of Evgpgration: Molecular dietillation differs from

ordinary distillation in that evaporation takes place from a quiet

.6-

surface. Since there ie no surrounding gas, there is no boiling

point. Distillation occurs to some degree whenever there is a tem-

perature difference between evaporator and eondeneer. An increase

in this temperature difference will increase the rate of distilla-

tion.

Langmuir's(3l)(lh) equation for the rate of evaporation of

metals in a Vacuum has been applied to rate calculatione in the die-

tillation cf organic substancee. where the Vacuum is sufficiently

high to allow the distance between the evaporator and condenser te

be small compared with the mean free path of the Vapor molecules,

the equatio for the number of molecules per secon, n, leaving the

surface is given by

n =

PAwhere

P = Vapor pressure in dynea per square centineter

A $ Area of distilling surface, eq. cm.

M = Molecular weightP

IE : Ideal gas constant

T = Temperature, °K

This equation has been applied extensively to the distillation of

organic substances by washburn(3h) and Burch$7) As can be seen,

substances having Values of the ratio of P to H nearly the same

cannot be separated to any great extent.

-7-

Fa¤cett(15) gives, as the ideal rate of evaporetion, the

equatien

u = 5.83xlD°2P\f:-2:-where

. H = Hase in grams evaporating per secend per squre

centimeter

H = Molecular weightlfT = Absolute temperature

l P = Saturation pressure in millimetere of mercury

In practice this rate is never reached, even at the low— pressurea, auch as (1 x.lOf? m. Hg.), which exist in the apparatue.

lf the temperature difference between the evaporator and cendenser

is sufficiently large to ineure th eondensaticn of all vaper unle-

culee leaving the eodenser, then any difference between the ideal

rate and actual rate of evaperation must be due to collieiens be·

tueen distilling melecules themselves, or between distilling‘mole•

cules and molecules of residual gas. These cellisions wuld then

cause some molecules to return to the surface of the evaporater.

Even by making the gap between evaporater and cendeneer amaller

no appreciable change in distillation rate is net•d.(15) In fact,

little difference ie obeerved ever a wide variation of gap dis-

tances, as long as the gap is within the mean free path cf the va-

por moleculee. l

·8-,

Vacuum ggguirementsz It is not necessary to exhaust com-·

pletely the apparatus used for·melecular distillation. The chief

aim is to reduce the number of collisions between distillate mele·

cules and.melecules of residual gas• Therefore, all that is nec•W

essary is to remove enough of the residual gas so that a uajority P

of the evaporating mclecules are allowed to pass unhindered. This

is equivalent to assuring that the distance between evaporator and

condenser is within the limits of the mean free path of the vapor V

melecules.

Bogligg Point: Since boiling point is defined as the tem-

perature at which the vapor pressure ef evaporating nclecules is

slightly greater than the pressure above the liquid, it is evident

that there is no boiling shnce there is no gas pressure above the

liquid. Evaporation takes place frem a quiet layer, and there is

no equilibriun between the vapor and the liquid on the evaoratingl

surface. The rate of evaporation increases with temperature, but A

under moleeular conditions, no boiling occurs.

Still Perfoggägge

Pressure: Pressure has no effect on the rate of evapora-

tion.(l5) A high pressure will restrict the passage of vapor mole-

cules to the condenser, the number of eollisions increasing with the

pressure. If the passage to the condenser is clear, that is, if the

vapor melecules are not hindered by the molecules of residual gas, a

.9- 9

further reduction in pressure has no effect. Ordinarily pressures

between l x lO°“ to l x.lD·7 mm. of mercury are used since little

or no hindrance is encountered.

The maintenance of suffieient vacuun is required since the

rate of distillation decreases with increase of pressure. Since a

leak•proof piece of equipment is practically i¤possible, some neans

of continuous removal of gas has to be provided. The quantity of

gas to be removed depends on the type of still being used. Usual

sources of gas are: _

l. Air leakage into the apparatus

2. Volatile vapor in feed

3. Decomposition products in feed

Air leakage into the apparatus can be reduced to a mininn

by careful sealing. Numerous designs of degassing equipsent for the

feed are known, usually the feed is degassed by the fore·pu¤p. De-

composition products can be controlled by controlling time of exe

posure. Serious troubles are encountered if even uinute quantities

of feed are decmposed. Maintenance of deeired pressure becomes

extremely difficult when deceposition occurs, causing the distilla·

tion being carried out to be worthless.

The Poiseuille expression for the rate of flow of a gas

through a duct does not hold at the low pressures eneountered.(15)

Unless connecting pipes are sufficiently large in dianter and short

in length to allow the vapor to pass without a pressure gradient,

.19-

high vacuum cannot be produced or maintained.Hick¤an(16) states

that the process is one of diffusiong that the passage of vapor out

of the system is dependent upon gas diffusio alone. This is based

on the work of Knudsen(29) who showed that the mass transferred at

low pressures is small in comparison with the volume, giving riss

again to large pipes and short connections. Fawcett(l5) reports

that making the gap between the evaporator and condenser too ssall

1 produces a pressure gradient, giving pressures in the still 100

times those in the pump.

Tggperature of the Distillandz The rate of distillation

increases with temperature. Distillations, therefore, should takee

place at maxinu temperatures. The limiting factor on the tepera·

ture of distillation is decomposition; as tmperatures are increased,

decomposition increases. Therefore, distillation should take place

at the maxinu temperature which gives no decomposition.

Careful control of temperature is necessary to eliminate

decomposition. This causes determination of evaporating tempera-

tures to be of prtme importance. The effective evaporation tem-

perature is the temperature of the outer layer of distilland, or

feed mixture. It is almost impossible to measue this, so it is

cuetemsryilé) to measure and control either the average temperature

of the film on the evaporator or the temperature of th heat source

t supplying the evaporator. Hick¤an(16) measures the temperature of

the residue leaving the evaporator surface. It is evident that the

.11-

temperature obtained is relative, no accurate temperature can be

obtained.

Fig; Thiekness:(l6) Attempts to decrease the thickness ofl

the film have probably done more to bring about the development of

molecular distillation than any other one factor. ‘äherc boiling

occurs, that is, under ebulliative conditions, the surface of th

liquid is continually renewed. As this condition is replaced by

evaporation frm a quiet surface, renewul of the surface is very

slow, csusing unwanted substances to be distill•d• Also, thermal _

deeomposition is affected greatly by the thickness of the distilland.

The rate of deconposition increases both with tmperature and with

time of exposure. Consequently thin files, neans of surface re-

newal, and minimum time of ezposure to heat source are imerative.

By keeping the filn shallow the desired constituent is allowed to

evaporate in a miniaun of time. 'all still designs are such that

the liquid surface is maintained as thin as possible with constant

regeneration of the surface.

A8 can be seen, when regeneration of the surface is not ac-

coplished, the efficiency of a still decreases considerably. A8

the partial pressure of the mcre volatile eonstituet decreases,

the gross rate of eyaporation decreases. Also, as the outer layer

becomes less concentrated in the more volatile conponent, less rola-l

tile components evaporate more rapidly, giving less separation.n

“ Thermal Efficigggz: Because of the proximity of the evap•

orator and condenser and because they necessarily face each other,

-12-

thermal efficiency is low. A heat balance determind by Faucett(l5)

gives heat utilized for distillation as 2% of total• To increase

efficiency two things can be done. One is to increase the tempera-

ture of the condenser relative to that of the evaporator• This has

a definite limit and is impracticable• If the condeneer is too hot,

reevaporation occurs at the condenser resulting in a loser distilla-

tion rate.

The second thing that can be done is to reduce the emmissivity

eoefficient of the evaporator. Polished baffle•(1b) have been used to

reflect part of the heat, but the path is obstructed if refletors are

hot, and if cold, no distillation occurs. Hicksan(l9) states that the

prise radiator is not necessarily the hot evaporator surface, but say

be the oil on its surface. with highly polished surfaces on both

evaporator and condenser Hickman found that the hot oil on the evap-

orator and the distillate on the condenser were the bodies receivin;

and emitting heat•

Hicknan(l6) states that it is possible to obtain a very thin

film on the surface of the evaporator and conenser by whirling the

at high speeds• This increases the thermal efficiency since the oil.

then ceases to emit rdiation• The little heat that ie esitted by

the polished surfaces is also reflected back•

-13-

Tzggs of Stille

Early attempts at molecular distillation failed because of A

lack of suitable equipment to produee necessary Vacuum. In 1916,

the invention of the eondensation pump by Langmuir(3o) allowed nee-

essary pressures to be obtained. üeveloment was relatively slow

until 1922, when Bronsted and HeVesy(3) devised a still for the

searation of the isotopes of mercury. Since 1922, considerable

work has been done on still designing. At present there are three

general types of molecular stille — the pot still, falling•film

still, and the centrifugal sti1l•

The Pot Still: The first workable still was that of

Bronsted and HeVeey§3)(k) as noted above. A pool of mercury wos

allowed to evaporate slowly in high Vacuum and was condensed on the

surface of a cool bulb held just above the evaporating mercury.

Additional work was done on the separation of the isotopes of

chlorine using similar apparatus.

Waterman and eo·workers(38) about the same time performed

distillations of coconut oil using a water cooled condenser.

Burch(8) did coneiderable work on the distillation of organie

material, introdueing a series of trays to agitate the distillad.

This was the first attempt to produce thinner layer: and better

stirring. Probably the transfer frau pot stille to falling·fLh¤

stille occurs with this introduction.

-1g.

The pot still ie similar to usual laboratory stills• A

flask containing the distillan is connected to a vertical con-

denser• As the distilland evaporates, the vapor molecules are con-

densed on the sides of the condeneer, usually cooled by air, and the

distillate is collected from gutters placed low and on the inside of

the condeneer• The vacuum drawoff is at the top of the condenser,

the outlet being of sufficient size to allow the necessary vacuu¤•

Falling Film Still: At the same time, Washburn,(36) Hick-

man,(2O) and Bupeh(8) were experinenting with new stille in order

to decrease thermal decmposition and give better separation• Eick-

man(20) is credited with the first falling film still and has done

considerable work on its development. In this type of still the die-

tilland is allowed to flow by gravity down the outside of a heated

column• The thicknees of the film depends on the rate of feed and

viscosity of the material being distillsd„ A condenser is placed

around the vertical evaporating column and within the limits of th

mean free path of the evaporating molecules. Methods of degassing

the feed are introduced as well as means of taking off distillate

and residue•

Stille of similar design have been introduced hy Detweiler(l3) ·

and 6aterman(39) giving modifications for the distribution of the

distilland on the evaporator• Detweiler recoumends a finely matted

surface if the flow is slow. Hickman(2l) introduced a spiral pattern

and a rotating distributor to prevet oil rivulets on the column,

e -15-

Ba¤gp¤£g(1) recommended a cascade divider• Hick¤n°s rotating die-

tributor appears to be best, but considerable rivulets continued to

form giving rise to thermal decomposition•

lndnstrially, the falling film type of still ie divided be-

tween glass and metal stille. Where observations ars necessary, glass

is used. Multiple units have been grouped in one container allowing

successive distillations. Burrows(9) gives detailed explanations of

different still modi£ications•

Centrifugal Still: Hick¤sn°s rotary distributor for the

falling film still lead him to the development of a completely ro-

tating et111.(23) A rotating evaporator is placed in an evacuatsd

container, Feed is introduced at the vertex of a cone·shaped evap·

orator and travels outward in a very thin film to the periphery•

The more volatile component evaporates and passes to the condenser,

which may be the dome of the enclosing container, The residue is

caught in gutters st the periphery of the evaporator. means are

provided for drawoff of both residue and dietillate•

At present stille are being used in sizes up to five feet

for the diameter of the evaporator•(26) ißonsiderable modification

has taken place since the original still was introduced, but the

general method is the same•

The evaporator is generally made of epun aluinum, mounted

on a long shaft held in place by a packed bearing. The condenser,

in its simplest form, is the ander side of a bell-jar• On the

larger stille separate condensers are incorporated which are water

cooled, er cooled by some other liquid. Baffled condensers en the

five foot stille have recently been introduced.

Cggpggison of Stgllg

Pot Type Still: At best the zolecular still is inefficient

in regards to separation,(l6) the separation of one “theoretical

plate“ being maximu. The pot type still does not even approach this.

Since distillation with a pot still is unavoidably a batch process,

the thickness of the layer of distilland is necessarily great. Re-

gadless of the decreass in distillatio temperature the time of ex-

posure to heat source is great, causing considerable decoposition.

means of regenerating the surface offer serious problms. This type

of still is confined to laboratory use.

s Fallipg Film Type Still: Considerable improvement is found

in the falling film type of still.(1b) Heans of control of the vari-

ables encountered in a molecular still increase the degree of separa-

tion as well as the rate of distillation. As has been pointed out,

several methods are emplcyed to give a uniform distribution of the

distilland on the evaporator column. The column tmperature can be

controlled either by using the vapor of a liquid of suitable boiling

point, or by resistance windings. Film thickness can be controlled

within limits by varying the rate of feed. Temperature of the column

and film thickness are greater than in the centrifugal still, but for pthe distillation of certain subetances, this condition is satisfactory.

„;7-

Qentrifugal Still: The centrifugal still offers the bestpossibilities for the improvement of efficiency.(l6) Time of exe

~ posure to heat source is decreased. By increasing speed of rotation

of the evaporator, time of contact is in the order of e fraction of

a second. Also, by controlling the rate of feed as well as the speed

of the evaporator, the thickness of the film is greatly reduced.Hick¤an(l6) gives a comparison that clarifies the situation. In a

pt still the thickness of distilland layer is 1 to 5 centimeters,—

tiee of expoeure to heat source is 1 to 5 hours. In the falling

film still the thickness is 0.1 to 3 mm., time of exposure is 10

seconds to 10 minutes. The centrifugal still reduces the thickneee

to 0.01 to 0.02 mm. and time of exposure to 0.0h to 0.08 seconds.

If ultra high speeds are used, it is anticipsted that the film canV

be reduced to 0.001 mm. and time of exposure to 0.001 seconds.

an early disadvantage of this type of still uns improper

distillate withdraxal. As distillation proceeded, less and less of

the more volatile components evaporated, giving side differences in

the quantity of each fraetion ebtained. As is the case in all molec—

ular stille, there ie no boiling, consequently means of determining

the best conditions to give maximum separation is of extrem i¤por—

tance. Hicknan developed the ”e1imination curve theory“ which has

· generally been accepted. Embree(lk) has worked out mathematical

derivetions of equations giving snalogous curves to those produced

by Hickman.

-18-

Eüimination Curve Theggg

In molecular dietillation the concentration of a substancein the material condensing is proportional to the partial pressure(P) divided by the square root of the molecular weight (M) of thesubstance• Therefore, substances having values of the P to thesquare root of H ratio which do not difter appreciably are impos-sible to separate by a single distillation• Where the identifica-

Vtion of constituents is desired, Hick¤an(18) has show that if themixture is exposed for evaporation for a certain time at each of a

series of temperatures, and the amount of a aubstance being con-

sidered is determined in each fraction, and these plotted againsttemperature at which the fractions have been distilled, a distilla·

tion curve is produced which he called an elimination curve. The

curve shows that since the vapor pressure is increasing, the amountsof the substance in each traction increases with tmperature until a

maximum is reached„ From the maximum, yields decrease rapidly be-cause the supply of the substance in the distilland becomes ex-Chausted• The maximum can usually bc located within a few degrecs•

Since boiling temperatures do not exist, these curves provide a(

valuable means of chemical analysis and are useful for studies con-

cerning multiple distillation.(lL) Considerable work has been done

in obtaining the elimination curve of fat—soluble vitamine usingdiaminoanthraquinone pilot dyes as indicators• By determining theelimination curve of a dye and knowing what cenponent distills off

.19,

with it, predictios, as well as identification: can be determined•

Distillabglitgs Heat is transferred to the dietillan uner

almost equilibrium conditions in molecular distillation• Therefore,

the rate of heat input (at constant temperature) and the shape of the

still are not ¤1gh1:1eant variables as they are in eullient distil•

lation• The rate of evaporation in.¤olecu1ar distillation can be

said to be proportional to the product of the nuber of molecules

and the probability that any nolecule uill distill• This probability,J

or tendency to dietill, has been called by E¤br•e(l&) the ”distilla—

bility”•

Distillatio from a Non Volatile Solvent at a Series og

lncreasggg Tempgrgturee: The derivation of equations giving close

approaimations to the analytical curves of hicknn have been carried

out by Embree•(lk) Such a curve is shown in Figure l, page 20, where

per cent yield in fraction is plotted against an arithmetic series of

abecissa representing the distillation tcmperatur•e• If the center

points of the chart are connected, an elimination curve in produced•

The temperature at which the naxheun in the elinination

curve occurs is a definite property of the substance, if the die-

tillatione are carried out so that the material ie distilld for the

n standard length of thme at a tmperature series o chosen that the

distillability increases by the sae factor at each higher tempera-

ture. The tmperature at which the distillation starts has no effecton the temperature of the maxhnun, as shown by Figure 2, page 21.

..„. ·. -..,„: en vr The actual eerieeef temperatures eheeen does net affeet the temperature ef the maxi-mun, ae lang aa the faeter ef increase in the distillability ia thesense in the eurvee 6@l3l•l"ld• The poeitien ef the elimlnatlon curve

ie markedly affeeted hy the length ef time during which the anb-

atance diatilla at each twperuture in the series, aa ehem by

Figure B: 9*8*£3• Theee curvee were caleulated fer the same aut-

etance and temperature eeriee ae in Figure 1, page 2.0. This ehewe

that lnereaaing the time of dietillatien shift: te the lßlllßä ef the

eliminatlen curve te a lower temperaturm wuaekenbaeh and tten-

hoekßz) imeetigated the effect of time on the elimination maximu

using celanthrene red und dibutyl aminoarzthraquinezxe. They feund

that in ueing a 20 minute cycle instead ef a 19 minute cycle ae

hieleaan did their eurvee ebend a nexiuwa 17 degreee lcner•°fet e., V 4., an. 4, e;e..;¤· 91;...,1;:,

rnb)Figure

le, Page lb, ehowe eurvee indicating effect ef distllling eubetaneee

with different vapor preeeurem eubetanee with the lover rate

ef change ef dietillability with temperature has a lower and flattereliminatien eurve. As has been peinted out, dietillablllty ia almost

preperticnal to the Vapor preaeurm Hehee, fer eubetaneea with a high

heat ef vaperieatien the distillability will lnereaee rapidly with

temperature und the elimination eurve will be herren end steep; sub--

atencee with a len heat ef vaporiaatien will have e breader andflatter eurvew

.25-

Distillate Drgigggcz The eliminstion curves given so farwere calculated for conditions which are difficult to duplicate ina real dietillation. Seldom does the volatile substance studieddistill over in the pure state, in fact, for the types of molecularstille that are used, it is desirable that a small amount ot solventdistill over with each traction to help remove the distillate fromthe conenser. Solvent here includes all the components present inthe solution being distilled, with the exception of the particularsubstance being studied. Controlled, or constant yield, oils havebeen produced which aid in the removal of the distillate.

If distillation is carried out at an arithmetic series oftemperatures instead of a series chosen so that the distillability

7

increase: by a constant factor at each temperature, the curve will6

have a slightly different shape, being lower and flatter than thecurve calculsted previously, as shown in Figure 5, page 26.

Distillation from a Non Volatile Solvent st a QontinuouslyIncrgggggg Tempg;atg;g=(lk) For sone distillations it is convenietto have the distillability increase continuously instead of by die-crete jmps. This situation may arise when a still is being suppliedmore power than is necessary to nsintain constant teperature. Al-though the ways in which the distillability can vary with thme are

infinite, the simplsst conditions will suffice to show the effect.Figure 6, page 27, shown this effect, the shitting of the temperatureof maxinum yield to a lower value when more time is taken.

.23.

Pregaration ggg Characteristics of Sgpthetic Constang Yiggg

_Q;gg: In the original experiments on noleeular distillation hy

Bureh,(7) Waehbur¤,(35) and haterman,(37) each worker pointed out

the difficulty of securing adequate drainage of the distillate from

the relatively large cndensing surface necesearil3‘eployed, The

change in the degree of overlap in the drainage from one fraction to

another is serious when the volune of the fractione very considerably

during distillation, Further, it is not known how a change of rate in

the distillation of one molecular species effects the rate of dietil-

letion of another speciee occurring in the sale oi1,(2) In a die-

tillation of Vitamin A with auch dilnent, the potency ie low, with

little diluent potency is high, and it is not certain whether the

variation and quantity of diluent effects the distillability of the

vitamin, Because of this,Hick¤an(lB) adjusted the eomposition of

fish oils during analyticel distillation so that the rate of distil·

lation of the diluent remained constent, This was done by the addi-tion of constant yield, or controlled yield, oils,

Congrolled Yield For effective use, such an oil

should satisfy rathäg strict conditions, It should distill over

the range 100 to 260*C, in a continuous molecular still, yielding

approxhnately the same weight of distillate fer each equal increment

of temperature, It is a convenienee if the distillates are liquid

at room temperature, and in order that the oil be related as closely

as possible to natural oil, it should consist of nixed glyceridee,

.29-

but should contain neither cholesterol nor free fatty acide.

A large selection of fatty acide ie not required for the

preparation of the desired glyeeride mixture. A blend of acetic,

butyric, the mixture of myristic and lauric acids from saponified

coconut oil, and the highly unsaturated C18 acids from saponified

perilla oil was foun by Baxter(2) to be sufficient for the purpose.

He also found that the elimination curves of vitamine an dyes are

best detenuhned with a conetant yield oil distilling at tempera-

turea from 100 to 26¤°C.

Auxilgagg Väguum Eguigggnt

Vacuu Pggpgz The production and naintenance of the high

vacuun required for nolecular dietillation presents special prohlens.

Pressuree in the rage tree l xl0°‘

to 1 x.1D°7 mm. of Hg. cannot be

ohtained by mchanical pups. Gaeses pumped include air and various

eubstances which exist in the vapr phase at these low pressuree,

some of which condense as the pressure ie increased touard atmos-

pheric. Such substances include water Vapor and products of decon-

poeition, and are labeled ae condensable contalinante.

Pumps used to evacuate the stille operate on the condensa-

tion principle an are designed to operate at pressuree on the in-

let side of from 0.01 mm. of Hg. down to the Vapor pressure of

their own working fluid. These pumpe are baeked by uechanical oil

filled vacuu pumpe which discharge to the air.

.39.

hercggg Condenaatigg Pggp: A8 has been pointed out previous-

ly, Langmuir(3O) developed the first condensation pump using mereury

aa the working fluid. In this pump the mercury is evaporated, the

resultant vapor is passed through a jet and then condensed, and the

liquid mercury is returned to the boiler. Mercury condenaation punpscan attain pressuree on the auction side of 0.001 mm. of Hg., this

being the vapor pressure of mercury at the temperature of the water

jacket.

Cold Trapp: For usual use this vacuum is ot sufficient.At first cold traps were incorporated in the vacuum line between the

condensation pump and the system to be evacuated to further redue

the pressure. High costs limited this to laboratory use. Conse-

quently work was done to develop low vapor pressure oils which could

be used in the pumpe.

Oil Coggensation Pggpz Bureh(6) developed several mineral

oil distillates possessing vapor prsssures st room temperature of the

order of l x.lO-5 to l x l0°6 m. of Hg. Use of these oils in con-

densation pumps proved their suitability. At the same time viscous

oila and greases were developed which ers useful in sealing joints.

The oils produced by Burch are in common use today and are sold under

the name of “Apiezo¤P oila. Hickman(2h) developed certain pure es-

ters that could be used, but these are not in geeral use. Greater

speeds are obtainable than in the mercury pmps because the molecular

sight of the oil is about 300 times that of mercury and the moleeulsr

U·Bl·

volume is about 15 times greatsr,(h1)

Operation of condensation pumpe using oils is very shuilar

to that of pumpe using mercury, The oil is hosted, producing va-T

pers which again are forced through nozzles, entraining diffusing

gas moleeules in the Vapor stream, The vapors are condensed, die-

charging the gas molecules to the forepump which exhausts them to

the amuosphere, Baffles are placed at the suction end of the pump

to minimise the number of liquid molecules diftusing back into the

system,

Fractgonatggg Cgggeggggion Pggp: where cohdensable vapors

are encountered, certain modifieations are necessary, Condensable

vapors which condense and dissolve in the pump fluid increase the

total Vapor pressure of the pumping liquid causing decrease in

attainable Vacuum, To eltninate this, the outlet pipe was heated,

condensable vapors passing into a water oooled cohdenser and being‘ trapped, Hickman developed what is known as a fractionating con-

densation pump, This pnp eosists of separate boilers arranged

so that the condensate from the boiler containing the liquid of

lowest Vapor pressure returns to an external boiler instead of to

g its original boiler, Here the more Volatile cnponents are sep-

arated out before the condensate is returned to its own boiler,

This is more or less a stage pump, the first, or fine boiler, con-

taining the lowest Vapor pressure liquid, Such a pup gives max-

imum Vacuum at the suction end, Three boilers are usually employed,

.;2-

plus the external boiler for separation of volatiln components,

Tube connections between boilers allow the working fluid to pass

from the external boiler to the boiler of highest Vapor pressure,

then to the boiler of intenusdiate Vapor pressure, and finally to

the finest boiler contining the low Vapor pressure fluid,

Vacuum hgasgggggtg: Provided that the pressure is low

enough, the performance of s.molceular still is independent of the

pressure, All that is necessary is such measuremsnt and control as

will assure sufficient Vacuum, giving approxinatc readings of still

pressure, There are several gauges in use today, none of which

give completely satisfactory results, the condensable Vapors present

in the still creating the problem, The hcLeon gauge cannot be used

at all, condensable vapors make it uselsss, The Pirani gauge is

used byHiek¤an,(l5)(16) but gives only miniuun of satisfaction,

Its range is fro 500 microns to l x„l0°h’¤m, of Hg, This gauge is

a covenient one to use, lt works on the prineiple that the heat‘ lost by the hot filsnsnt is a function of the pressure, It is

usually calibrated in air and the assuuption is made that the cali-

brstion can be applied directly when the gauge is used on a ¤olsc•

ular still, Faweett(l5) has shown that readings 2 to 5 times too

high are obtained, He also pointed out that after s short time or-

ganic vapore collect on or around the tube, giving extrenely in-

accurate readings, Hiekmsn replaces and recslibrates hie gauges

often to minimise this conition,

.33,.

Types of Vacuum Gaggeg: The ionisation gauge is inaccurate

in the presence of crganic vapors and gives readings that are less

accurate than those of the Pirani gauge, Its range is from 1 110.3

mm, of Hg, to 1 1 10.7 mm, of Hg, The Knudsen gauge, which is ex-

tremely accurate and delicate, requires too careful attention to be

of auch value, Fawcett(15) recoamends for everyday use a simple elec· · A

tric discharge tube, Since it is only necessary to know that suf-

ficient vacuum is maintained, a discharge would indicate still pres·

eure has reached too high a value, As long as no diseharge is evi·

dent, still pressure is assueed to be maintained,

111ck¤g¤(1b) uses a tilted U··tub• aanometer filled with amyl

phthalate (density equals 1 at 25°C,) as a check on his Pirani gauge,

It is used as a double deflection instrument by placing one lieb in

connection with the vacuum to be measured and the other limb is con·

nected to s. reference vacuum, If the connections are reversed

and the difference between the two readings is halved, the true

‘ pressure is obtained, Absolute preesures dem to 10 uicrons are

obtainable,

Future of Molggular Distillatgon

Two main disadvantages limit the canmercial use of noleculsr

distillation equip¤ent,(l6) The process is ixmerently a heat com-

sumer, thermal efficiency being about Ä, Also, the fraetionating

power is small, approximately that of one **theoretical plete“ , For

.3;.

these reasons, in its present develoment at least, it cannot cou-pete with ordinary distillations. It ie confined to those processesable to bear the high process costs, such as the vitamin industry.

with advanees in heat economy a wide field ie open for itsdevelopment. The natural olle and waxes which form one of the w¤r1d'sgreatest resources of organie raw material fall within the scope ofthe process. Plastieisers, plastics, resins, polynere, sugar deriva-tives, rubber intermediates, and petroleum residues are a few of thematerials that can be purified by nolecular distillation.(16) alreadythe vitamin industry has made considerable headny. In other indus-tries where quantities of thenmal liable substances are being wnsted,solecular distillation is making headwmy. The developnent of mole-ular distillation is just beginning. lndustrially, the greatest de-velepment lieg ahesd,(l6) The centrifugal still offers a means ofrapid advancement of the molecular distillation process.

III. TAL

Äe EQQQSBotIn

order that future studies ot all factors involved in theaolecular distillation process may be carried out, a moleeular still

having esse of adaptation to changes uns necessary. It was the pur··

pose ot this investigation to design and cenetruct a molecular still

that would enable these studies to be nade.

B. Plan og lg}stiggtg

l. Qgew of the Ligraturgz A review ot the literature

was made to obtain the latest developments in the field of moleeular

distillation. Particular attention was given to still design, still

characteristics, and still construction. g

2. Prelininary drauings of the rctors,bearing, packing nut, roter plate and shatt, base plate, collectiug

gutters, and heating unit were made and submitted to Gcmas (Zigarette

Machine Ccapany, Salen, Virginia, tor nachining.

3. Construgtion Q Qggggtr Equipnent, including a true,

rotors, roter plate and shatt, collecting guttere, heating unit, base

plate, bearing, packing nat, and necessary pmps md notor assublies

was designed, construeted, and assembled tor use in this investigation.

.35-

L. ägperimental Determinations: The apparatus was checked

for leaks and the vacuum gauges checked for accuracy. Two test runs

were made to determine whether the still operated. A mdzture of 30

weight Essolube metor oil and Celanthrene Red dye was used as the

feed mixture. Data taken was in the form of observations, rather

than quantitative date, enabling recommendatins for improvements to

be made.

5. Evaluatggn of Results: The errors made and the problems

enountered in the assembly of the apparatus, elinination of leaks,

and the operating procedure of the still were discussed. The results

obtained in the experinental tests were evaluatd and ineorprated in

the discussion of the results. Becoumendations for future work were

also discussed.

The following materils were euployed in this investigation:

Qglggthrene Red Qge: Coeetreted, 125%. Used es pilot dye

in the feed mixture. Obtained from E. I. duPont de Nemeurs and

Company, Hilmington, Delaware.

Qgrasic»Cement §Insu•lugg}: Adhesive Cnent Paste Ho. l.

Used as ceramic bining material in the hosting unit. Obteined fr!

. Fisher Scientific Company, Pittsburgh, Pennsylxania.

.37..

ggggg: Red Enuel 21,}.1. Used for sealing joints and con·-

nections. Obtained from General Electric Corporation, Schenectady,

Heu York.No. 11,-636. High Vacuum Grease. Used

for eealing joint: and connections. Obtsined freu Fisher Scientific

Ooupan, Pittsburgh, Pennsylvania.Gg_e”_e gelgsceneb Heavy Vacuum Grease. Lot Ho. SAL Used

for eealing joint: and connections. ütained freu Distillaticn Pred-

uctc, Incorporated, Rocheater, Heu Iork.

High Vacum Grease. Effective

at pressuree of 10.6 n. of Hg. Usefnl fron-&0‘C. to over »200‘O.

Used for sealing stopcecks and connections. Obtained Iron Dow Corning

Corporation, Hidlam, Michigan.g__|1_l §§so]_‘_¤l_gg Ho. Q}: Used in feed nixture. Obtained freu

Speedway Esso Station, Blackeburg, Virginia, in quart csns.

¤· aamata E

The following equipnent and apparatus were uaployed in this

investigation:0

§till : Figur lk, page 55, shows assubled still head,

the parts ef which are listed below:

gsee Plsgg: (One) Figure 10, P~9·|• ls'7. Made of steel.

Diaeeterx 15 in., thickness: 1/2 in. Drilled to accmedate jvacuun outlet, feed inlet, residus outlet, distillate outlet,

..33..

heater wire inlet, vacuzm gauge inlet. Used ae support

for still assembly. Machined by Comas Cigarette Machine

Company, Salem, Virginia.

Bearjängz (One) Figure 9, Paße L,6. Made of steel. '

Diameter: 2 in., length: 6 in., drilled to fit a 1/2 in.

shaft. Bearing surface is l/8 in. oil ilpregnated bronze.

Faetened to base plate by preseed fit joint. Used as bear-

ing for shaft and as support of still assembly. Machined

by Comes Cigarette Machine Company, Selen, Virginia.

Bgargqg Sugggg: Made of cast iron. Used to mount

bearing and base plate aseembly on the freue. Obtained fra!

Comes Cigarette Machine Company, Salem, Virginia.

Collggtigg Gutters: (Four) Figure 8, page 1+5.

Made of rolled bronze. Dianeterz 3/h in. Constructed to

fit around the periphery of the four rotore with a 1/32 in.

elearance. Each has an outlet of 1/I, in. copper tubing 1/2

in. long. Used as collecting gutters to collect residue

coming off the rotors. Maehined by Case Cigarette Machine

Company, Salem, Virginia.

Heatigg Qgt Sheg: (One) Figure 12, P¤&• 1+9.

Made of aluminuu. Diaueterz 5 in., height: 1/2 in. Holloved

out to wall thicknese of 1/8 in. Hole drilled in center

slightly oversize for 1/2 in. ehaft. Used ae heating element

for evaporators. Machined by Comas (Zigarette Machine Compaq,

.39-

Salem, Virginia:

(Plastallic) Garlock 926: Gonsiste of

pure asbestos fibre, shreds of soft metal, graphite, and

heat resisting lubricant: Used aa paeking in packing nut:

Obtained from Garlock Packing Company, Palayra, New York:

Paggg Hug: (One) Made of steel: Diameter: 2 in:,

length: 1: in: Threaded to fit bearing: Used es housing for

packing material around shaft: Machined by Coaas Cigarette

Machine Gcmpany, Salem, Virginia:§•_;_tLgg: (One) Figure 7, page Ah: Made oteast

aluainum: Bianeterz 5 1:1:, thickness: 1:169 1:1:, sngle ot

inclination: 75°, distance from Vertex to periphery: 2:59

in: Used as an evaporator: Haehined by Gemas Gigarette

Machine Gmxpany, Salem, Virginia:_§_g_t_L•g£: (Che) Figure 7, page /+l+• lade of cast

Disaster: lr:79 in:, thickness; l:&9l in:, angle

of inclinatiom 67}**, distance {rm Vertex to periphery:

2:59 in: Used as an evaporator: Machined by Gauss Gig-6

arette Machine Company, Salsa, Virginia:

gg§g;! (Ode) Figure 7, page Mw Made ot eastF

aluninm: Disaster: £,:l:8 in., thicknessz 1:795 1:::, angle

of inelinaticn: 60°, distance Iron Vertex to periphery:

2:59 in: Used as an evaporator: Machined by Comes Gig-

arette Machine Company, Salem, Virginia:

.;,9-

gggg: (One) Figure 7, page M, Made ef east

alrminum, Dianeterz l,,ll in,, thicknees: 2,075 in, , angle

of lncllnatien: 52}*, distance frcm vertex te periphery:

2,59 in, Used as an evaperater, Hachined hy Cases Cig·

arette Machine Compazw, Salem, Virginia,

Roter Flag gg Shag: (the) Figure 11, page h8•

Made ef steel, Diaueter ef plate: 5 in,, thiclmess: 1/I, ain,

Shaft was turned dem from same piece te diaueter ef } in,

Length ef shaft is 13 in, Used as support ter reters and asA

means er retating rotere, Hachined by Cenas Cigarette Ma-

chine Company, Selm, Virginia,l

§_gggg_¤__§E_g: Figure 15, page 51, shows the eenstruction ef

the vacuun eystm, the parts ef which are listed belcv,

gp: (One) Cence-liegavac Pup with neter, 220

velts, 325 RPM, capacity ef free air per ninute: 31 liters,

Used as ferepuup in vacuum systm and as degassing pump fer

feed, Obtained Iren Central Scientific Compamr, Chicage,

Illinois,§_@_p: (One) Glass, three stage, eil cendensatien

pump, Type GF-25 w, Capacity: 25 liters per sec, at 10%

mm, ef Hg, 90 te 250 wette, 65 te 108 velta, Used te ob-

tain low preesures within the still, ßtained freu Die-

tillatien Products, Incerperated, Recheater, New York,

..;,1..

Vacnun Ga__ug: (One) Prelininary vacum gauge.‘

Scale from 500 nicrons to 0.5 nicrons. Hot wire thermo-

couple type. Used to determine pressnree within the still

within its range. Obtained from War Surplne. Hanufaeturer

unknown.

Vaguum Gänge: (One) Ionisation gauge. Range from

1 x10,-l°

ma. of Hg. to 10.8 mn. of Hg. Used as neans of

determining extremely low preesures. ätained fron War y

Surplus. (Hanufactured by Philamonic Radio Corporation,6

ue.; rsrk, N. I.) ·

Auxiligq Agparatne: Figure 15, page 51, shows the auxiliary

apparstus, the parts of which are listed below.

Aspirator Bottle: (One) Six liters capacity, mabber

tubing connection. Used as feed reservoir. Obtained from

Fisher Scientific Ccupany, Pittsburgh, Pennsylvania.

_ Bell•jar: (One) Low form, udth knob. Height inside:

6 in., dianeter inside: 10 in. Used es cover for still head

and as condenser for the distillate. Obtained from Fisher

Scientific Cgpgny, Pittsburg, Pennsylvania.

Glass Tubgp (8 um., 10 nm., 3/Ä in.) Soft glass.

Used to connect various pieces of apparatns. Ubtained fra

p Phipps and Bird, lncorporated, Richmond, Virginia.gggt_g_r_: (One) Variable speed A.C. Weetinghouse

ADS, 1/L HP at 10,000 RPM, 115 volts, 60 cycles, 50 degree

..;,2-

C. rise, centinueus operation. Style 957656-B, Serial IR.

Used te drive shaft and roter aesenbly. Übtained free

Westingheuse Electric Corporation, Pittsburgh, Pennsylvania.· Petentiemetegz (One) Leeds and Herthrep twpera-

ture indicating petenticneter. Serial He. 89571. Cold

junction cenpensater, direct reading. Used tc determine

the temperature ef the residue leaving the evaperater.

Obtained from Leeds and Nerthrup Company, Philadelphia,

Pennsylvania.

(One) Used to determine RPM ofmeter. Obtained frau Fisher Scientific Company, Pittsburgh,

Pennsylvania.

Tgsla geg: (One) Vacum tester, high frequency.

·Used te test for leake. Obtsined fren Fisher Scientific

Company, Pittsburgh, Pennsylvania. '

Used te make connections betweenglass jeints, netel jeints, and glass te netal joint:.

Obtained frau Chenical Engineering Departlent, Virginia

Pelytechnic Institute, Blacksburg, Virginda.

Side _g_Ar;" Egg, Rubber Stegggge, snd ellaneggg

Used ae distillate and residue reeervoirs, cen-nections fer feed system, and neszle fer feed entrance. Ob-

tained freu Chwical Engineering Departnent, Virginia Poly-

technic Institute, Blacksburg, Virginia.

·-h3-

@$13: (100 ft.) 0hromel—A. Free freu iron, 18gauge, 0.l.06 cha per ft. Used as resistance wiring in the

g heating unit. Obtained from Fisher Scientific Company,

Pittsburgh, Pennsylvania.

läge (6 tt.) 1ron··Consta.ntan. Leeds and Northrup

inenlated thermoconple wire. Used to deternine twperature6 E

of reeidue leaving roter. Ottained Iron Leeds and Northrup

Conpany, Philadelphia, Pennsylvania.

¤·Generg

The equipnsent, a five inch oentrifugal aolecular still, was

designed and constmcted as shown in Figure 15, page 51. Each inte-

gral part was assemhled separately, such that any one part could be

removed from the assemhly without disturbing the other parts. This

allowed freedom of neintenancc and was ot value when changes were

necessary.

Fgame

The frame, the details of which are shown in Figure 13, page

52, was made of one inch angle iron. The lower section of the frme

was obtained from War Surplus and was 30 inchee square on the top end

44.

2.075"

----1T ...... ...... ....... ...... 1.49"

~ r

y- ........................ ...........

......

' I I

~----------~r·~·-~--------~J~

4.79" ---->-l"' l I

I I

Material of Cona~~ction - Caat Aluminum

Revision VIRGINIA POLYTE.-HNIC I N.:J TITUTE DZPARTHENT OF CHEMICAL EN ;rlNEERlNG

BLACK3BURG. VIRGINIA

HOTOR3

Drawn by : 1B-i Checked by: 'f£9 Approved by :~

Date: 6-1 Scale: 1/2' : 1" Figure No: 7

45.

''rant Vio\"1

- - -t-1 I I

-..~ ido '!i·;);;

. '

Sootion Vio\1 A ~

,'. ote : ~ach Hot<;>r }~as a r,u ttor r.c:J.chinod t(· :'i t o.round tho pori~hory Hith 1/52 i n . cl~iuw.nce •

. nter L1.l. of .Jon t.rw-:: tion- ·'r onzo

I

·'

Havl a .i.o

' . I

XLL ·;; ~u:.z

Jrl1wn . by : i t,;; , Ohock od by:'f!d £J ',~ i>r o ved by :.:lltfr~

'1 t o : 6-1 .,..· 10 . ' 1 1/ ') ·- l t~ .. CU. I - -

Figure Nd :8

41,.

Top View

r· .. ~ r."1 I I

J/1 I I I I

r- I ,_j I I I

I I I

6''

I I

I I I

I

I I

Vi

011 t.rl.lat 11d t. \

Mat&rial of Construction:

Hild Steel

tubes were silver .

soldered 1n place.

47.

Top View

to 0.5000•

fl•ont View

·1 e visinr VlnGL:I .);; rt~., ., ... :t;T c ... -- ~ ... ' !J.~L ~i,.riN~:..; nm

l.J r 'fffi byt fiJ_.i. \Jheoked by: ~l.f-1,1)roved by;

Jatar 6-1;)-48 ., n - -J ..)C£1 10~ -a - 1

~ igure No : lO

•

4-8.

I

j I

-~--4- --~ I

I

Top View

Front View

Material of Construction - Mild Steel

A - Aligrn;1ont Pegs,

B - Drilloo and tappod for 10-'2 machine screw

Reviaion VIRGINIA POLYTCCHNIC IN3TITUTE DEPARTMENT OF CH~<' !CAL ENGlllEL:IUNG

Bt.AOKJ BURG VIRG 1 •

ROTOR PLAT r~ AND SH.A.t< r

Drawn byz Datet 15-48 Oheoked byz o/~~ Ooale: 1j2•:1a Approved fiy :.7WTF Figur No : 11

~ 5" . ' --I

jb------ --e _: ~ -------- ]£ '. w I I

fRCNT VIE';/

terial of Gonatruotion - ~·t Aluminum ·

on VIRG~NIA POLY.TC.:'JliTi!G I H..iTITUT•; DEPA~TMGNT OF 0}\8M!GAL ~;GIN~~;:miNG

H~ T:::R SH!~LL

Dra-wn ··by: . Jato: Ohockod by:~ . J calo: i ' : 1" Approved ~. by:~ f igure Ho: 12

,59-

button and 31 inches high. The frac was strengthened hy weldingA

four hraces to the uprights 16 inches above the base.

Fged Bottlg Suppggg: Additionl height was needed to have

the feed introduced hy gravity flow. Prelininary experiuents ehowed

that gravity flow could be ohtained with a head of eleven ische:.

an additional support was added to the frame to give this head.

Figue 13, part F, P¤8¢ 52,show• the details of the con-

struction. One inch angle iron uns used. The height of the freue

was increasd eighteen inhes. The feed inlet in the base plate was

six inches above the original franc, thns s head of twelve inches

was allowd for the feed.

ggäggggjägggtz Boiler plate, l/L in. thick, was used tosupport the still head, held in place hy the hearing. Figure 13,

part D, page 52, shows the details of the construction. A fourteen

inch by five inch section of the boiler plate was welded to the

channel iron. Four holes, one·half inch in diaaeter, were drilled

in the support. Te bearing support, uade of cast iron,1¤as bolted

to the hoiler plate with four, three·eighths inh nachine holte.

Ionigggggg Gsgge Sgpggggs The top of the freue was used to

support the ionization gauge. Figure 13, part Eh page 52, shoes the

details of the constructio. One inch angle iron was used. The

construction was sch that the base of the ionizetio gauge would

reet on one face of the angle iron, the other face served t keep the

gauge in place.

•

Legend

A - Still Head. Figur~ 14 show~ details. B - Fr ame. F igure 13 shows details. C- Diffusion Pllmp. 0 - Forepump. E - tonizetion Gaug e .

F- Fead Oontafner. G- M o tor.

H- Rt$fdue 6 Distillate Containers.

F

..

SIDE VIEW

•.

~B

~..-______ 3 o•:..._• -----~~

3 o" I I I : ., I I I

1'--tc _ l I I I ' I ,--r - ~, I I I I I I I I

I \ II I • I I I 1 'r£!4+1• I ,It ~·,r.r , , 1 1 I : IJ: :: I I : Ill 1 ~ 1 1 1

I ·~lfi'!,' , I I I '- - - '1h•'-- - - L

L--- --1- - L j - - -- I- - - - l -~-~

TOP VIEW

. I I g B o : I

I II E I II

•• •I II II II

,,

B

4 8"

'

FRON T VIEW

Revisions VIRGINIA POLYTECHNIC INSTITUTE ,., DEPARTMENT OF CHEMICAl. ENGINEERING

BLACKSBURG VIRGINIA

STILL orawn by : j 6J,

Trace ·d by: ~J. it PP roved by:~

ASSEMBLY

0 at •: lo - I~ - 48 I If t / 1

scalt; fz :.. l - 0

Figure No: 15'

c j

'·

0 0

A B c D

E

F

E

~

0 0

D 1

Legend

Motor Supports.

Diffusion Pump Sup ports. Distillate a Residue Support.

searing Support.

Ionization Gauge Support.

Feed Bottle Support.

F

11... 1

l : I

\~

.~ :

~/

SIDE VIEW

3011

I

I .. --,;. I

I I

I I

3011

I I

~ ~

I I I

I I.! .. II I _:i._

I I c~

TOP VIEW

r '"" fJt J~

I I

j -rr····--- ---- -------. f • _t_f• •• • ••

... , t II A

ll I .t-

B

FRONT VIEW

I

4911

L

Revisions, \ VIRGINIA POLYTECHNIC INSTITUTE

DEPARTMENT OF CHEMICAL ENGtN£d­BLACKSBURG_1 VIRGINIA

FRAME ,

0 r o w n b y : ~ D a t e : b - 15 -- 4~

Traced by : ~ Scale: 1" = 1'

Checked by:~ File No:

Approved by;~. I Figure No: 13

-53-

Qgggusiog snap Suppgrt: Figure 13, part B, page 52, show

p the details of construction of the diffusion puhp support. One inch

flat irc was usd. One support was placed seven lnches from the

front of the frame, the other twenty and a half inches from the font.

The supports hung three inches below the supports to which they were

welded. This allowed the lntake of th php, hen in position, to be

two inches below the vaeuuh outlet in the base plate.

Motor Sugpgrt: One inch flat iron was used as supports fer

the motor. Figure 13, part A, page 52, shows the details of the co-

struction. One support was placed fourtcen inches freu the front of

the frame, the other twnty seven inches. This allowed enough space

between the ntor and the supports to obtain vertical alignment of the

shaft•

Residue an Distgllggg Bottle Sgppgggsz One half inch angle

iron was used as the platfona for the residue and distillate bottles.

Figue 13, part A, page 52, shows the details of the construction.

The iron extended in front of the frame ten inches. It was bolted to

the frame with l/L inch steve bolts.

‘ éasßlt§t;llHgggz The still head was asshbled as shon by Figure

lt, page 55. The bearing was inserted in the bearing support and

fastened in place with a set sorew. This plaeed the base plate one-

half inch in front of the franc. (The base plate and bearing were

-5g-

joned by a pressed fit connection by Comes Cigarette Machine Company

before delivery•)

The heating unit was then placed in position• The base of

the unit eas one inch from the surface of the base plate. It was

held in place by three 1/k inch screen, passed through the holes in

the supporting lege and into the base plate. Metal eashers insertcd

between the lege and the base plate allowed the screen to be tightened

and the desired distance from the base plate to beobtained.The

rotor plate and shaft was then passed through the heat-

ing unit, through the bearing, and held in place hy the surface f

the bearing. A collar, placed between the base plate and the heating

unit, kept the ehaft freu aoving inward.

The packing nut as then ecrewed on by hand until th s1ight·

est pressure eas felt on the shaft. (This operation will be diecussed

in detail in a later paragraph. The shaft and packing were turned doen

to a mirror finish before the rest of the apparatus was aseeubled.)

The second collar eas then placed on the shaft to arodd outward move-

ment of the shaft. A febre eaeher was later placd between the collar

and t end of the packing nut to decrsase friction.

The rotor was then placed on the roter plate ad faetened in

place. The gutter correepoding to the roter used was placed around

the roter and fastened securely; care eas taken to properly center the

gutter with respect to the rotor•

II

II t -------~

E A s

F G H

1511

1-----........ ----.r-4,--- -- --- --------. I I --- --------~ ------- ~ --~..J

----------- ----------~ I I

~------------~~·-- -------------1

PLAN V IEW

8 c 0

A B c 0 E

F G H

L egend

Gutter. Fig ure a sho ws details. Rotor. F1gure 7shows details. Rotor Plate. Figurttl l shows dtta il•· Heating Unit. Figure 12 shows details. Base Plate. Figure 10 show• deta ils.

Bearing. Figure 9 shows detai ls. Packi ng Nut.

Collar.

R•visions VIRG IN lA POL V'TECHNIC INSTITUTE DEPARTMENT OF CHEMICAL ENGINEE RING

BLACKSBURG, VIRGINIA

STILL , HEAD ocawn by: 't ·

. T r a c e d by: ~ri~Jl Daft : 6 - If- 48

I o /'' scale: ~ ~

Approved by:~ Figure No: 14-

..55..

e Tygon tubing was used to connect the gutter outlet with the

residue outlet in the base plate, This was later repleeed with copper

tubing, The feed nezzle, ten nillimeter glass tubing, was bent end

placed eo that the tip ot the nozele was just above the vertex of the

roter, Tygon tubing was used te hold the nozsle in place. A burettecletp, fastened to one brace of the gutter and to the nossle, eeeured

the nossle.§_g_t_e_g: The netor was nounted on the supports provided for it

on the true, One·-eighth inch round head steve bolts held the noterin place, Hetal washers ineerted between the nete: base and the sup-

ports allowed vertical slignnh. A two inch length of 1/k" insidediueter, 3/A" inch outside diaaeter, heavy·-walled rubber tubing ns

ueed te connect the notor shaft to the roter shaft,

Difrugion E: The diffusien puep wae plaeed in the bracketsTprovided for it en the frame. lt was not eecured to the supports,

Glaes tubing, 20 m, in diueter, was used to join the intake end ofthe pup to the vacuun outlet in the base plate, Tygon tubing we:

need te nake the connections, The joint: were eealed with glyptsl,Rubber tubing, 3/8 inch in diaeeter, was used to connect the cooling

water inlet with the water supply, The cooling water outlet was een-

nected to the drain with rubber tubing of the eene dianeter,

Foreg: The forepunp, or llegavac punp, was placed on thebaseetthefrsne suchthet the inlettotheforepuepwasinlineudth

the outlet ef the diftnsion pm, Glase tubing, 20 u, in dieeeter,connected the two pupe, Tygen tubing was placed over the ends of the

-57-

glass tubing and the joints were sealed with glyptal, An oil seal

was nade at th intake of the forepunp,

Fged Systens A six liter aspirator bottle was used as the

feed reservoir, It was placed on the top of the franc an connected

to the feed inlet in the base plate with glass tubing, 8 , in di-

ameter, A stopcock was placed in the li four inches from the inlet

to the base plate to control the rate of flow into the still,

A 250 cc, separatory tunnel was placd in one hole of a two

holed rubber stopper, This was placed in the top of the feed con·

tainer, The bottle was connected into the vacunn line with glass

tnbing, 8 un, in dianeter, One end of the tubing was inserted in

the other hole of the two hole stopper and the other end was sealed

with a glass seal to the line from the dittusion punp to the fore-

puup, This allowed the feed container to be evacuated by the tore-

punp alne, giving a pressure differential between the feed container

and the inside of the still, The connections were made with tygon

tubing and sealed with glyptal,

Residue and Distillate §gstensz A section of plywood, 3/h in,

thick, uns placed on the protruding platform in front ot the frame,

The residue bottle, e $0 cc, side arm flask, and the distillate bot·

tle, a lO cc, side arm tlask, were placd side by side on the plat-

form, The side arm of the residue bottle was connected to the reeidue

outlet in the base place with glass tuhing, 10 mn, in diameter, Glass

tubing,A6

m, in diameter, connected the side arm of the distillate

-53.

8bottle with the distillate outlet in the base plate. an 8 mm. stop-

cock was inserted in a one hole stopper an this placed in the residue

bottle. A solid rubber stopper was placed in the distillate flask.

All connections were made with tygon tubing and the joints eealed with

glyptal.Theggggogple wgggz The fibre insulated, iron•constantan,

Leeds and Northrup thermocouple wire, previouely ealibrated by Dr.

F. J. Gradishar, was checked and inserted into a hole punchd in te

Vacuum line to the diffueio pmp. Since this was the largest hole

learing the base plate, less interference would be encountered here

than elsewhere. The wire ae brought into the still and placed in the

’ tygon tubing, and later the eopper tubing, one inch from the residue

outlet in the gutter. The eonecting tubing was larger than the out-I

let in the base plate to awoid the possibility of interferene to

flow of the residu out of the still. The holes in the tubing were

eealed with glyptal.

It was later found that leakage occurrd between the insu-

latio and the wires. The insulation was removed for a twe inh

length just as it entered the still. The bare wiree were then eoated

with glyptal. A thin coat of high Vacuum grease was applied after

the glyptal had dried.

Heatggg Unit wire: Rubber insulated wire was used to con-

nect the resistance wie of the heating unit with the source of

electricity. The ineulation was removed from the end of the wires

-59-

and they were brought into the still through the hole provided in

the base plate. The two wire: were connected to the resistance

wires of the heater. The inlet tube was eealed with glyptal. The

other ed of the wire was connected to the terminal: of a owerstat.

Vggugg Gggges: The ionization gauge was placed in the sup-

ports provided for it on top of the frame. The prelimiary gauge,

or hot wire gauge, was placed beside the feed container.

The vacuun tube: used to measure the vacuum were connected‘

to the vacuum outlet provided for it in the base plate with glass

tubing, 10 mm. in diamcter. The socket ends of the tubes were in-

serted in holes drilled in a section of pine board. The board was

fastened to the trame with wood ecrews. Tygon tubing was used to

make the connections ad the joints were sealed wdth glyptal.·

Powggstags: The three powerstats wre placed on a table n

beside the frame. Connections were made from one powerstat to the

motor, one to the heater: of the diffusion psp, and one to the heat-

ing unit in the still.

Potentiometer: The temperature measuring potentioucter was

connected to the thermocouple wire and set on zero position.

Bel1•jg;: The bel1·jar wes placed against the base plate in

auch a manner that the distillate outlet tube was just above the

bottom of the inside of the jar. It was sealed with high vacuum

grease. A wire cord attached to the frame and to the knob on the jar

held it firmly in place in case of vacum failure.

.59-

§ggg;gg_§gt: It was early realized that the rotating shaft

was a cause ef considerable concern in obtaining the Vacuum. Con-

aequently, auch tine ws given to the turning down of the packing

and the shaft to acquire a.¤irror finish on each.

After several types of paeking were tried with no success,

Garlock packing 926, Plastallic, was eployed. Packing, consisting

ef fibre asbestos, shreds of soft metal, graphite, and a heat re-

aisting lubricant, was plaeed in the packing Hüte The packlng was

ef two forms, loose and rolled. The nut was filled to within three-

quarters of an inch of the end with loose packing.A A one•half inch

ring of the rolled packing was inserted in the end. The shaft was

introduced into the packing and the nut placed on the bearing and

tightened by hand util a alight pressure was felt on the ahaft.

The collars were then tightened and the shaft connected to the motor

with the heavy walled tubing. The motor was startd and the shaft

allowed to turn at an average speed of 150 HBH for fifteen mintes,

or until the pressure en the shaft had diminiehed. The packing nut

was then tightened by hand again until a slight pressure was felt n

the shaft. This procedure was continued until the nut could be

tightened no more. Each time the nut was tightened, the motor was

stopped to keep the shaft {rum binding.

•6l·

The nut was then taken off and additional packing addd.

The same procedure was followod ae when the nut was first packed, a

small amount of loose packing was added followed by a ring of rolled

packing. The shaft was again inserted in the packing and the nut

tightened on the bearing. This procedure was followed until four

additional packings had been canpleted. Leakage through the shaft

and paeking was stoppd as the nut was being tightened the fifth

time.

Heating Uggts The heating unit for the evaporator was con-

struoted in the following anner. Twenty·seven feet of number

eighteen chromel wire were tightly and evenly wrappsd around a

welding rod, l/8 inch in diameter to fonm the heating element proper.

A wooden frame, fitting into the hollowed out aluminum shall ma-

ehined for the heating unit, was made. A pattern of coneentrio

ciroles was made on one side of the frame with nails, and the ooiled

wire placed around the nails following the pattern made. A ceramie

fire elay binder was placed on the bottom of the shell ad the wire

lowered on it. Additional ceramic material was forced around the

wire. Care was taken to keep the wires from toching the shell. An

oversize shaft kept the hole in the heater shell from being plugged.

The frame was preseed into the shell until the wires wre slightly

below the top edge of the heater shell. The whole asseubly was

allowed to dry for forty·eight hours. After drying, the wooden

block was removed and the unit was examined to assure separation ef

-62· I

the wire and metal. The shaft was removed, leaving a hcle in the

center. A thin cost of insn·lute cement paste was spread over the

tcp of the unit and smoothed with a spatula level with the top of

the heater shell. The unit was then baked at ll5'F. for forty-

eight hours.

F. Methods of Pgggedggg

The following methods of procedure were employed in the

operation of the apparatus during this investigation:

l. A feed mixture consisting of two quarts of Essolube

Motor Oil, ö.a.E. 30, and 0.0l0& grams of celanthrene

red pilot dye was made.

2. The megavac pump was turned on.

3. with all stcpcocks closed except the one in the top of

the residue bottle, the bell-jar was plaeed against the

base plate, care was taken to insure that the bell-jar

was placed so that the distillste draw—0ff tube in the

base plate was immediately above the lower inside sur·

face of the bell-jar.

L. The stopcock in the residue bottle was closed. This

caused the system to be placed uner vacuom, the

pressure thus holding the bell·jar firmly in place.

(The bell·jar was further securd by tying it to the

frame with a wire cord.)

-63-'

S. High vaeu grease was presaed around the lip of the

be1l·jar where it rested against the base plate. Care

was taken to insure that a eontinuous seal was made.

6. After a ten minute interval the system was tested with

the tesla eoil to deternine the degree of vaeuum ob-

taind. If the ionized gas within the system exhibited

a pink eolor, the pressure was low enough to turn on the

preliminary vaeum gauge.

7. If the gauge indicated more pressure than 200 microns,

the tesla eoil was used to find probable leaks. If

leaks were found, they were sealed with glyptal or

high vaeuum grease, depending upon the location of theU leak.

8. Ehen the pressure gauge indieated 0.mierons or less,

g the cooling water to the diffusion psp was turned on.

9. The powerstat for the diffueion pump was then turned o

and the voltage set at 30 volts to allow wermup and

degassing of the oil.

10. After twenty minutes the voltage to the diffusion pump

was inereased to 90 volts, the voltage fond to give

maxi! pumping speed.

ll. The mixed feed was then introdueed into the feed bottle

by filling the separatory funnel and very slowly a11ow~

ing the mixture to enter the bottle by regulating the

.5g-

etopcoek of the funnel. In this manner the feed uns

degaesed, no recirculatien of the feed was possible for

the degaasing prcess.

12. The powerstat to the heating unit was then turned n,

AO velts being set on the instrunent. The heater was

allowed to werm.up and degas for at least foty·¤inut•s.

(It was feund that a considerahle lese in Vacuum occurred

when the heater uns turned on. After appreximately fort:

minutes, the Vacuum could again he obtained.)

13. When the vaeun gauge indicated 0.5 nieren: or less, thev

system was ready fo operation. The fed stopcock was

then opened slightly allowdng the feed to enter the inlet

nozzle at apprexintely 10 ce. per minute.

la. when the feed mixture reaehed the roter, the powerstat to

the meter was turned on and the speed of the motor set at

approximetel: 2000 HPI, a setting of 80 velts on he power•

stat.

15. The feed rate ue: then adjusted until Lt appeared that an

extremely thin filn of oil covered the retor. (It ua:

sonetines necessary to sdjust the speed of the note

also.)16. as the residue hegan to come off, a temperature reading

uns taken hy neans of the temperature indieating potentie-

meter and the input voltage to the heater was adjusted

·65·

accordingly. In the tue runs made, a setting of 30

volts on the powerstat would give a tenperature high

enough for dietillation to occur.

17. The Teeder counter was used to deternine the RPM of

the motor.

18. when the readings rmained constant, no changes were

mac in the settings unless different conditions were

desired•

19. The data taken coneisted of (l) tina, (2) temperature

of the residus leaving the roter. (3) BR of the motor,

(R) preesue in the still. Beadißsß Here taken at ten

minute intervals. ·

20. The separation of te dye fra the oil was detenuined by

visual means. The inside surface of th bell—jar, ating(

as the condenser, was eolored red. lf the hell-jar b-8

came wann, a fan uns placed in a position to cool the

surface of the b•11·jar. If the red color did not then

appear, it uns nceasary to increase the voltage to the

heater.

21. When it was seen that aeparation had occurred, the feed

uns cut off, the motor uns stopped, and the current to

the heater and to the diffusion pn uns cut off.

22. After the diffusion punp had cooled, the cooling water

‘ was turned off.

.66..

23. The megavac pznp was stopped, the b•1l···,jar untied, and

the stopcoek in the residue bottle opened. The bell-

jar was then removed. ’

ZA. The apparatus was then cleaned and made ready for theT

next run.

G. Data gd Hesgg

The data and results obtained in this investigation were

in the form of observations. During the construction and operation

01* the apparatus, the errors made and the information gained relative

to improvements were n0ted„ Three runs were made to test the appa-

ratus, all of 1'orty minutes duration at a temperature of 78°C. and a

pressure of less than 1 x 10* mm. 01* Hg. with the EFH of the notor

set st l8®.U

The observations made during the course of this investigation

are listed below

1. Exeesslve vibratioa occurred, originating in the megavaopump.

2. The rollere on the frame had a tendency to transfer the

vibration to all parts ot the apparatus.

3, A mechanieal means of holding the bel1·jar in place ‘

would be of value.

I;. Extreme care must be exercised in obtaining a. mirror1'inish on the paeking and shaft.

-67-

5, At the temperature maintained during the runs made,78°C,, ne charring occurred on the surface of theroter,

6, The distillate collected in e pool on the bottom of thebell•jar and drainage into the rceiver was extremelyslow,

7, The introduction of the feed inte the evacueted feedbottle was secoupenied with vielent bubbling, causingthe oil to pass into the Vacuum line an thence to theforepup,

8, With the feed bottle at a greater pressure than therest of the system, the flow of feed into the still weseatisfactory, The flow could be cotrolled with thestopceck,

9, Becirculatio of the feed though the still would be ofgreat advantage, both in the degassing process and inobtaining samples for eliuination curve date,

10, Tygon tubing ie useless in the presence of oil and heat,ll, It appeared that the Celle·eeal grease used distilled at

the low pressures obtained, even thogh the vapor pres-sure as given on the label was lower than the pressuresobtained in the still,

12, A thin coat of glyptal made a better seal than a thiekcoat, or blob,

•68·

13, Oil er grease of any kind caused the shaft to bindafter a mirror finish had been obtained,

lk, The motor was found to heat excessively ir the s1ight·est pressure was exnrted on the shaft by the tighteingof the packing nut,

15, The oil in the diffusien punp was found to containforeign matter,

16, Leak finding with the Teela coil was difficult becauseof the large amount of metal present,

A 17, Reinforcement of tagen tubing was necessary at pressures

below 500 mdcrens, 418, A loss in Vacuum occurred when the heater uns turned on,

The pressure increased fro¤.lees than 5 x l0°*‘m¤„ ofHg, to AOC micrene, requiring A0 lintee of operatienof the heater to regain the Vacuun,

19, Distillatio was seen to occur at a·te¤perature of 78°C,

and a pressure less than 5 :.10-&'¤¤, of Hg, (The needlewas off of the scale of the hot wire gauge, The insideof the bel1•jar was seen to be colored red,)

20, A Vacuun line, separate and distinct freu the liquid line,

was needed for both the residue and distlllate receivere,21, The inlet tubes in the base plate were found to leak,22, Leakagc occurred between the insulation and the wires of

the thermocoule leads,

•69•

23• Containers connected to the Vacuum system with 10 mm.

diameter tubing evacuated quicker than containers cen-

nected with 6 m. diameter tubing•

Zt. Coneiderable heat was produced because ot friction be-

tween the end of the nut and the ¢ollar•

2S• The base plate, b•l1•jar, and inlet tubes in the base

plate beeane uam during operatiom

.79-

IV. DISCUSSIOH '

A. Discussion of Egsulta ang Rggggggggatiggg

Qgggral: In this investigation the data and results were

obtained from the design, construction, and operation of a five inch

centrifugal nolecular still. Two runs were made to deteruine whether

the still would operete. The data obtained from these runs was in

the form of observations made relative to the inprovement of the appa-

rstus.

Since the work done in this investigation was of a construc-

tion nature, the experience gained was of iuportance for those con-

tinuing this work. In the following paragraphs the observations made

an the experience gsined will be discuseed in detail. Recummendations

resulting fra these observations will also be diecussed.

Vgbration Encountegggz Considerable vibration occurred be-

cause of the forepunp. The pump, located on the base of the frame,

caused exeessive vibration in every piece of the apparatus. It was

found on one occasion that the filauent in one of the vaeuuu tubes,

the hot wire tube, had parted. It ie possible that the tube was

burned out, but nore than likely it wos broken by the vibration since

care was exercised in the use of the tube.

It usa found quite difficult to naintsin leak proof seals.

The constant motion of the whole assembly caused frequent breaks in

the glyptal sealing cenpound. These leaks were extremely suall and

+7l·requiredconsiderable time to locate.

The alignment of the shsft and the motor was also effected

hy the vibration. After several hours operation of the pump, lt was

found that the motor had shifted from its center position. This showed

up by the motor overheating. As soon es the alignment uns again node,

the temperature of the motor decreased. The shitting of the motor took

place even though it was securely fastened to the traue and lock uashee

were used. The constant vibration losehed the nuts, and on two oe-‘

casions, the nuts dropped off."

The greass seal around the lip of the bell·jar eas affected hy

the vibration. It was found that this seal around the lip of the bell-

jar, where it fitted against the base plate, demanded a continuous seal

of the high vacuum grease. The eonstant Vibration caused small breaks

to occur in the grosse allouing air to enter the still.

The ollers on the bottom of the frame, although they caused

no vibratio, allowed the vibration produced by the pop to be more

readily transferred to all parts of the ¤PP&ratue.

Recomendatigggz It is recommended that the following be

done to eliminate excessive vibration of the apparatus:

1. The forepump should be located on a platform, or mount,

separate from the rest of the apparatus.{A

eoncrete

mount, placed directly behind the apparatus, would serve

as a stable platform for the pop. It is possible to set

the pump on the concrete floor of the lsboratory. This

.72-

would serve the same purpose as a mount.

2. The rollers on the bottom ef the frame should bs re-moved. In their place should be installed adjustable A

lege, giving both balance and stability to the frame.-3. The Vacuum tubes should be installed in a shock•proof

nount. It is possible that the vacuua reading on thegauge wos affected by the Vibration ef the tube.

Pgggggg Nut: As has been pointd out, the packlng and pack-

ing nat was an important part of the vacuua rystem. Considerable thne

was necessary to turn the ehaft and packing down to s.airror tinish.At present the shaft and paeking have been turned don to suh a fin—

ish that leaksge around the shatt does not occur, er it it does occur,ls so small that the vacuun of less than 5 x.lD°*'¤¤. of Hg. could beobtained.

A

It was observed that even the slighteet pressure exerted by

the packing on the shaft would cause the shaft to bind. The shaft

anat nova freely in the paeking. As the speed of the shaft was in-

creased, heat was produed in the packing nut. The expansion hat

thus occurred was enough to increase the pressure on the ahaft. After

a mirror finish was obtained, it was found that aore vaeuu could beobtained with the shaft rotating than when it was still. The saall

increase in pressure of the packing resulting fr¤¤.the expansion of

the packing produced a better seal around the shatt.

.73..

It was also noted that considerable frietion occurred at

the end of the packing nut where the collar was located. This frie-

tion caused a considerable rise in the temperature of the packing

nut, resulting in a slight binding of the shaft. This in turn

caused the motor to overheat. A fibre washer placed between the

collar and the packing nut eliminated this trouble. No such condi-

tion was found at the other collar resting against the base plate.

It appeared that the shaft had a tendency to nove oatward, that is,

away from the motor. This was probably caused by the rubber coupling

connecting the shaft to the notor.

p It was thought that after a nirror finish had been acquired,

one drop of a light weight oil would effectively seal the shaft. This

was not the case. When an increase invaeuun was noted during the

turning down of the shaft, one drop of light weight nschine oil was

pleeed on the shaft where it entered the packing net. Imediately gthe shaft stopped, the motor being overloaded. Although the binding

was found to be very slight, it was necessary to back the packing off

a fraetion of a turn to release the shaft. The notor was then turned

on and the operation took place es before. After several hours of op-

eration of the apparatus, oil was again placed at the end of the pack-

ing nat. The sans result occurred. Further investigation indicated

that no oil seal was necessary around the shaft, the lubricant present

in the paeking was sufficient to seal the shsft.

.7g-

goungationz The packing nut played auch an important

part in the construction of the apparatus that the procedure followed

in sssling the shaft ie of value. lt is recommended that the following

procedure be used to pack the nut and turn the shaft down to a mirror

finieh.

1. The packing should be put in the nut by hand and packed

by hand around the shaft. Ho packing with a blunt in-

etruent should be attempted. when placed on the bearing,

the nu shuld be alowly turned hy hand until the slight-

est pressure is felt on the shaft. The shaft should be

rotated by hand as the nut is being tightened to deter-

mine when the binding occurs. When the pressure is first

felt, the collars should be tightene and the notor allowed

to slouly turn the shaft. After a fe'minutes, two or

three, the speed of the motor should be increased to

about 2000 RPM and allowed to run at this speed for fif-

teen minutes to one half an hour. The notor shold then

be stopped and the shaft rotated hy hand to determine

that free movement occurs. The nut should then be

tightened by had until the slightest pressure is felt

on the shaft. This procedure should be repeated until

the seal is made. (It uns found in this investigation

that the nut could be tightened to its full threaded

length four tines, on the fifth tdne the desired finish

.75-

was ebzuusa.)

2. No oil, or other lubricant not already present in the

packing, should be placed in the packing nut.

3. After the mirror finish is obtained, all horizontal

movement of the shaft should be avoided, decreasing

the possibility of disturbing the seal.

Distgllgtg Qgauofgs It was anticipated that the distillate

drawoff would be slow. This was found to be the case. As condensa• .

tie took place on the inside of the bell—jar, mall streams, er riv-

ulets, were seen running down the inside surface. The distillate, in-

stead of running out of the still, collected in a pool on the bottol

of the bell·jar. This could have been caused by one of three things,

the hole was too small, the bell-jar slanted downward from the dis-

tillate outlet in the base plate, or the pressure in the distillate

bottle was greater than that in the still. The last cause wdll be

diecussed in a later paragraph.

Although the volume of distillate eating off at any one time

was small, it is possible that drainage to the outlet was so slow that

until the streams were fonned, he distillate did not reach the out-

let. lf the distillate had flowed to the outlet as it came off, the

hole in the base plate would have been of sufficient eine to accommo-

date it. As it was, he distillate arrived at the outlet in relative-

ly large quantities and could have blocked the entrance to the hole.

Since the vacurm takeoff to the receiver bottle uns through the liquid

.75-

line, it is possible that the distillate was held at the mouth of

the tube because no outlet for the air within the bottle was provided.

It eas believed that this uns the cause of the trouble, since after e

pool of considerable sise had fonued, no distillate passed through the

tube.

At first it was thought that slanting the frame upward would

inrease the drainage rate, thus elhninating the resistance to the

distillate drawoff. It was then realized that this would only in-

crease the qusntity of liquid reaching the outlet at one time, causing

the blocking of the outlet sooner. A pop could be euployed to take

the distillate off as fast as it was forued, but this was not believed

to be feasible. Nor can it be stated that tubing of larger dianeter ie

necessary, ths present tubing would probably be large enough if the

distillate could be drained off ae fast ae it is foneed.

Vgguun Lines to Distillate and Residue Regeivegez As was

pointed ot above, it was possible that the lack of distillate dra¤—

off could be accredited to a larger pressure existing in the receiver

than in the still. It was believed that this was the case. The pres-

ent construction had os line fra each of the receiver bottles, dis-

tillate and residue. At the start of a run these bottles were evan-

uated through these lines from the inside of the still, no direct lines

to the vacuum pups were present. As the run progressed and either

residue or distillate came off, the pressure within the receivers began

to rise because the liquid was still warm and 8111 quantities were

.77-

still evaporating. Since this gas was not allowd to pass off, a

pressure was built up in the receivers that was greater than the

pressure existing within the still proper. If such a pressure did

build up, the flow of liquid into the receivere would be stopped.

hggomehgatihx The following rec·~«endation is therefore

made to eliainate the resistance to flow enountered in te dis-

tillste and residue drawff tubes.

1. The receiver bottles should have separate connections

to the pumping aystau. This would allow the vacuua to

be maintained at the same, or a loer pressure, than

that existing within the still. Flos into the re•

eeivers would be added. A trap between the reoeivers

and the papa would be necessary to keep Iron contaaina·

ting the puaps with vaporise material tree the still.

lt is possible that oly residual gas, air, would be

discharged into he puaps, but at the taaperstures of

distillation that wold be needed, this is not belisved

to be the_case•

hgggggulahgon og the Fggg

There are several points to be considered relative o the

recirculation of the feed. It would afford the best possible means

for degassing the feed; it would allow the feed rate to be set at any

predetermined rate of flow·with a,ainis¤ of time and trobleg it

would allow continuous operation of the still without the breaking of

-73-

seale an the loeing of vacuum; it would elhninate the resistance to

flow encountered in the receivers; it would allow the number of stop-

cocks, which are a source of leaks, to be reduced; and it would allow

data necessary for the determination of elimination curves to be ta-

ken.

Dggaesgggz Proper degassing of the feed was found to be a

najor problem. In the first run it was foun that degassing couldbe accomplished faster by the slow introduction of the feed into the

previously evacuated reservoir. As the feed entered the reservoir,

violent bubbling, or foaming, occurred, filling the entire container.By the time the feed had all been introduced, degassing had been ac-

complished. Ho trouble was encountered in the feed systm during the

first two runs.

However, during the third run serious trouble occurred. Thefeed, already in the feed bottle, was degassed with the forepump

three and ons half hours before the beginning of the run. When the

etopcock was opened to allow the feed to enter the still, bubbling

occurred to auch an extent that it was impossible to regulate the

flow. At the tip of the feed nozzle the air·feed nizture frothed

over the evaporator, prohibiting distillation and causing contanina-

tion of the product.

It was noted that after four hours of degassing by the fore-

pump, considerable bubbling occurred in the feed reservoir when th

teela coll was placed against the side of the bottle. It appeared

-79-

that the feed was degsssing, no evidence of decomposition of the

feed was fouml•

ömall quantities of the feed mixture were seen to pass

through the vacuum line to the forepunp when the feed was introduced

into te degassed feed bottle. This was elininated by slewing down

the rate at which the feed was introduced. However, a trap betwen

the feed bottle and the pump wold eliminste an possible chance of

coutaminating the pump with materials used as feed•

äggommeggatiogx On the basis of the difficulty encountered

in the degassing process, the following recomendation in made.

1. A magnstic pump should be installed that would take the '

material from the reeidue container and return it to

the feed reservoir• Degassing of the feed could then

be aecomplished in the easiest manner• If auch a pump

were installed, the feed could be slowly run through

the system, the heater to the evaporutor being off•

As the feed entered the still, it would be distributed

hy the rotor and degassing would oecur• The feed, of

the same composition as that which entered, would then

be pumped back to the feed reservoir• Succeseive cir-

culations of the feed in this msnner would thoroughly

allow degassing to take place.

-39-

Advggtaggs og a Rggirgglgtion Sgpggx In addition to the

above advantage given for a feed eirculating system, three other

advantages were evident.

1. The installation of a magnetic pup for the feed would

eliminate the resiatance to the flow of the residue into

the residue receiver. As the reeidue passed out of te

still, it would immediately be puped back to the feed

reservoir. A continuous flow would therefore be main·

tained and aid to flow, rather than resiatance to flow,

would take place.

2. lt is believed that auch a pump would eliminate the nec-

essity of a separate vacuun line to the residue receiver.1

The outlet provided hy the pump would allow the residue

to pass unhindered into the container, blocking of the

inlet would be eliminated. The pressure within the still

y and within the reservoir would be the same. Furthermore,

an increase in the feed rate would not cause additioal

resistance to flow of residue into the container. The

pump, in order to give the increased rate of flow, would

also have to return the residue to the feed reservoir atthe same rate. Consequently, an increase in the flow rate

of the feed would automatically demand an increase in the

y flow rate of the residue back to the feed container. The

rate of flow into the residue bottle would be dependent

.31-

upon the capacity of the pump and the dianeter of tubing

used to connect the pieces of equipment, not on the in-

let tube to the residue receiver.

3. A magnetis pump would allow the data necessary for the

determination of elindnation curves to be obtained.

This advantage will be discussed in the following para-

graph.

Sampling: Although no analyses were made in this investiga-

tion, it is easily seen that such analysen would require the breaklng

of seals to the residue and distillate containers. As the temperature

of the evaporator was increased, fractions of different compositions

would sone off. As each different fraction was celleted, the system

would have to shut down, the scale broken, the pups stopped, the

heatere cut off, and the samples removed. Then the snnections would

have to be remade, the vacuum again obtained, and the heaters turned s

on and regulated to the proper temperature. when the next sample was

collected, the same procedure would result. Most of the time would

therefore be wasted in collecting the sample and getting the apparatus

back into operatio. This could be avoided by the installation of a

pump to recireulate the feed. For elinination curve date it is not

nceesary to obtain a sample of the residue. The punp would therefore

continue to recireulate the residue and the saeplee of distillate could

be wdthdrawn, if some system was incorporated that would allow eamples

to be taken at will.

.32-

Recommendation: On the basis of the above discussion, the

following recommendation is made.

l. A system for obtaining snap samples of distillate

should be installed. Figure lb, page SA, shows such a

system. Two small round bottom flasks are connected hy

a glass joint. A stopcock, placed between the two flasks

acting as a check valve, could be opened or closed to

allow material to flow from the top bottle to the bottom

bottle or to stop the flow, as the case might be. When

enough distillate had collected in the top bottle for

sampling, the stopcock could be opened allowing the sample

to flow into the bottom bottle. The stopcock could thenf

be closed, sealing the bottles from each other. The stp·

cock to the diffusion pump could then be shut and the

sample witdrawn from the bottom of the flask by mans of

the stopcock providd. When the sample had been reoved,

the stopcock in the bottom could be shut, the stopcock to

the diffusion pump opened. As the distillate was coll•ct·

ing in the top flask, the bottom flask would be evacuating.

In this manner, samples could be obtained et will. The

three stopcocks are additional sources of leaks, but these

can be sealed, and the advantage of such a system would

more than compensate for the extra stopcoeks.

-g3-

Feed Rate: As a fifth advantage for the installation of amagnetic pump, the control of the rate of feed could be added. Thepresent construction allowed different rates of flow, but no pre-determined rate cold be set. By determining how·mmch feed hadpassed through the still in a measured time, the average rate offlow could be determined. This condition would have to be correctedbefore elimination curve data could be obtained.

In determining still characteristics and elimination curvedata, it would be necessary to run the feed through the still at adefinite temperature and definite flow rate, draw off samples, andthen run the residue, as feed, back through the still at an increasedteperature of distillation and the same rate of flow. Since the

properties of the residue would be different after each run, the flowrate would have to be adjusted to give the same amount of feed perunit time.

With a circulating pump of the magnetic type, predeterminedflow rates wold be possible. The motor driving the pump could beof the variable speed type. Calibration of the pump would then allowpredetermined speeds to be employed. With a variable resister in theline, a chart of inet voltage against pumping speed could be made.Calibratio eurves of materials of different properties could be madeallowing freedm of operation. 1

5*4.

To StillAE

—•— ToDiffwleiom•

YDietillate (but

Revision VIRGINIA PGLY’I‘JCHNIC II·Ii$I‘I„TUTE.DSPARTMEZNT OF CHEMICAL TSNGINEZLLRING

BLACKSBURG VIRG "I ;

DIJTILLATS .3.^.I«LE’LlI»§G LSYISTELLEJravm by: jdn?Date:Checkedbyzqßg Scale: NoneApproved by: « ·‘¢

Figmxro No: 16

-85-„

&éggg of the Hgatg; tigt

During a trial run it was observed that a. loss of vacuunoccurred ae soon as the current to the heater was turned on. Thepressure at the time was considerably less then 5 x l¤°" ¤. of Hg.The powerstat connected to the heater was set at fo;-ty volts. In•

mediately the vacuun decrsased, after two minutes the indicatedvacuun was &00 nicrons, a considerahle increase in pressure. The

powerstat was turned off and within three ninutes the pressuredecreased to less than 5 x 16-k m. of Hg. Severel attmpts toregulate the heater voltage without a lose of vacuue occurring weresade; all were failuree. The Voltage was then allowed to reusin on

T

tor longer periods ot tine. After fort; ninutes, the pressure wasseen to drop. A »

It is not kamm whether the loss was due to the evoluation ofTgas entredned in the heater, er to decuposition of the ceranie eeeentused in the hester. It appeared that the lose wee due to the eVelu··tion of gas, air, held by the materials composing the heater. Eschtiue the heater was expoeed to the atmosphere, the same pressure losswas encountered when the still was put into operation. The reedingeon the gauge were the same, a drop fra less then 5 x 104* nn. of Hg.to LOO uierons. The heater use exemined and no apparent change took

place.

. ·86-

Regonmeudatggz It is therefore recounended that con-

siderable time be allowed for degassing the heater. The

g powerstat to the heater should be turned to forty volts

as soon as the forepunp is started and the qstem put

under vacuun. This will allow sufficient tiue for theheater to degas since thirty to forty minutes are needed

to get the still in operation. It should also he rena-

bered that a greater tine will be required to obtain

sutficient vacuua to turn on the diffuaion pump.

Heat Tgansfeg gran the Haag;

ggg; Tgansgeg to the Bgge Plate: At the tine the first run

was being nade, it was noted that the base plate and all connections

to the base plate became warn, graduallv getting werner end warmer.

At no time did the parts become too hot to touch.

There is a poesibility that warping would occur if ezneeasive

heat was transferred to the base plate from the heater. It was learnedthat when the inlet tubes were soldered to the base plate by Comes Cig-

arette Machine Couparq, the base plate warped and had to be turned dom

the second tine. Although the tenperature obtained in this investiga-tion, 78*C., was insufficient to cause mrping, the tmperaturee nee-

easary for future distillations could be high enough to have some effect.

gomendationz lt is reconmended that sone means be provided

for cooling the base plate. It was found in this investiga-tion that a fan bloving on the assembly satisfactorily de-

.37-

creased the temperature of the base plate and aided in the

cooling ot the conenser ae well.

§£;ect of Heat gg thg Sgglggg hgteriglg: It was found that

the glyptal eeala remained undamaged by the heat transferred fro the

heater to the base plate and thence to the seele. There waa ne evi-

dence of the glyptal's softening or beeouing porous•

The Cello•seal grease, however, holding th seal between the

bell—jar and the base plate eas aff•cted• It was neticed that he

greaee began to sotten and ran along the lip of the bell·jar• On

the inside of the bell~jar small drope of a clear liqudd were eeen•‘ The nly possible source from which these drope ef liquid cold have

come was the sealing grease. It was believed that at the temperature

attained hy the base plate and at the pressure existing within te

still at the time, less than 5 x;lD·“'mm• of Hg•, distillation of

the grease took plane. It was ossible that sole entrained material

in the grease, auch es moisture, distilled and the grease was not

affected• However, eoftening did occur•

The Cello-Seal grease was replaced with the ßelvacee hamy

vacaaa grosse and it was neted that this greaae was not aftected by

the heat.

Eegougndaggons It is recoemened that the Gelvacene high

vacum greaee be used for eealing all joints that cannot be

sealed with glyptal, auch ae stopeocke an the lip ot the

bell-jar,

-88-

Charrigg of the Feed on the Evaggratogz In this investi-

gation no charring of the feed occurred on the evaperator. It wasnot assuued that charring would not occur. The length of the runs

in this investigation and the temperature of the distillation were

insufficient to cause charring. The longeet run made was fcrtyminutes, the highest tuperature of the residue leaving the evapo-

rator was 78%}., and the feed uisture was composed of S.A.E. 30

notor oil and Celanthrene red dye. The temperature of distillation

of the oil was lower than that of materiale used in cmercial no-

leculsr stills; therefore, ao couparison can be node.

TUse og th; Tggla-00;,1;: The Tesla coil was found to be very

satisfactory for locating leaks where glass or rubber connections

were made. Where metal forsed one part of the connection, the Teela

coll was uaeless• when the uetal was approached by the coil, srciug_ A

occurred between the end of the coll and the netal, elininatixg iteusefulness as a leak finder. Arcing also occurred through glass

where metal was beneath the glass. The leaks located by the Teela

coil showed up as bright spots on the outside surface of the con-

nection. Arcing was evldent through the leak into the system. Only

leaks of a certain size could be located in this umher; if the leeks

were too large or too small, other methods had to be used•

*89*

Igglation hethgg: The locating of leaks aroud netal joints

was difficult. A very volatile naterial, auch as bensene, gasoline,

or toluene was held around the joint and the tesla coll placed in a

position to cause ionisation of the gas inside the systen downstreen

frau the suspected leak. The difference in the color of the ionized

gas indieated the presence of a leak in that area. This method was

used often but on only one occasion proved to he of value. The size

of the leak controlled the amount of the material drawn in. Most of

the leaks encountered in this investigation were too mall to be de-

tected in this manner.

If a leak could not be located by any cf the above methods,

g the trial and error method, or isolation nethod, was used. The area.of the suspected leak was isolated from the rest of the vacuum systen

and the tesla coil used to observe changes in the color of the ionised

gas. Rapid changes indicated leakage, since the pressure within the

isolated section was changing. If a section was found to contain a

leak, all the connections in that eection were clesned and resealed

with glyptal or high vacum grease.

On all occasions the leaks were located hy one of the above

methods. The most difficult leak to find was located in the thermo-

couple wire entering the still. Since considerable netal was present,

the isolation method was used. It was finally fomd that the leak was

through the base plate. After considerable time eas spent trying to

locate the exact position of the leak, it was realized that the wire

..9g..

entering the base plate could be the source of the leak. Vacuus

grease was applied to the lead wires and an increase in vacuum was

imediately noticed. The wires were then skinned for a distance ot

two inches from the point ot entranee into the base plate and sealed

with glyptal, The vacuua increased considerably, Iran 200 microns to

more than 5 x 104 m. of Hg•

ggwazendaüonz It is reconnended that all wires entering

the vaouum system be skinned and carefully eealed• Al-

though one type of insulation was Iound to be leak··prooi',

the possibility ot leaks warrants this csnticcn

’Tygen tubigg was found te be very satistactory as a means of

making connections between glass and glass, glass and metal, and metal

and metal, A peeuliar property ot tygon increased its ettectiveness

c¤nsiderab1y• shere connections were ende with tygon tubing, as the

Vacuum inereased, the tubing begsn to eonpress around the joint re-

sulting in an almost air·tight seal. The conpression was so great

that ditficulty was encountered when removing the tub1¤g• It was

necessary to insert aus object into the end of the tubing which

allowed air to enter• If this was not dene, it was almost inpossible '

to raore the tubing without tearing it• Other types ot tubing tried

did not show this property• A

.91-

Tygon tubing, however, cannot ba used inside the still where

the pressure is low and the temperature high• It was foud that

tygon tuing armecting the residue gutter with the base place out-

let, became soft at a tnperature of 78°C. and a pressure less than

5 x.l0°h m. of Hg. The tubing had to be replaced with copper tubi¤g•

The tygon tubing used in the Vacuum line collapsed easily•

All connections of tygon tubing had to be reenforced to aroid partial

closing of the joint. In all cases this was done by inserting glasstubing of slightly snaller dianeter than the connection into the

space within the conection. This proved to bc satisfactony•‘ Qggommendationz Tygon tuning should not be used for uaking

connections if it will be in contact with natsrials above78°C•

Evacugtion Tg!

It was found that containers conected into the Vacuum system

with large diameter tubing cvacuated more rapidly than containers con-

nected with amaller dianeter tubing, even though the container was a

larger size or a greater distance from the pu¤ps• The residue bottle,

a 500 cc• flask, was connected with 10 nn. tubing• The distillatebottle, a 100 ce. flask, was connected with 6 un. tubing„ On all runs

it was noticed that the residue bottle showed more vacuun than the die-

tillate bottle. This was seen hy the color of the ionised gas when

the tesla coil was brought near. When the Vacun gauge lndicatd apressure less than 5 x 10% m. of Hg., the color of the gas in the

-92-

residue bottle was blue, the color of the gas in the dietillatebottle was pink. According to Iarwood(“l) blue indicated more vac-uun than pink.

It was thought that a leak existed in the distillate bottle.Considerable care ws exercised in sealing the stopper and connec-tions. No change in the color of the ionizd gas was seen.Afterthe

still had been in operation for an hour, the indicated vacuumin both bottles was the same. On the next run, the same observa-tion was made, a longer time was needed to evacuate the distillatebottle than was needed to evaeuate the residue bottle.

Qggoenations When making connections into a vacuunline, the largest possible tubing should be used. Anyone section ot a vacuun syste that uns connected intothe system with small diameter tubing wnld increase thetime needed to evacuate the system.

Seeling og Joint;

The obtaining of leak•proof jointe was an ilportant part ofthis investigation. Although leakage was never conpletcly stopped,it was rduced to such an extent that pressure: lower than 5 x;l0°‘nm. of Hg. were obtained. As was previously pointed out, tygontubing uns used for.naking connections wherever pessible. In thisinvestigation connections made with tygon tuning were sealed withglyptal, a thin coat being applied. The glyptsl was alloed to drytor twenty·four hours before the connection was plaeed uder vacuwn.

.93-

It was found that tuenty·four hours was sufficient tine for theglyptal to dry. It was also noted that a thin film of glyptal nade

a better seal than a thick coat, or blub. Qu several occasions itwas found that where a large amount of glyptal was present, thesurface only was dry, the interior was soft and fluid. These cn-nectione proved hy testing to he leaking. When the excess glyptalwas reuoved, the connection cleaned, and a thin coat of glyptalapplied, teste proved that leakage hd stopped. lt appeared that _ —

the undried glyptsl beneath the surface allowed the fonnation ofsell air holen. Although the holea were extremely shall, leakageoccurred to auch an extent that sufficient vucunn cold not be

ob·”\

tained.

For seals of a very temporery nature, high vacuun greasewas used ad found to e satisfactory. Where iumediat sealing was

deanded, high vacuua grease was used. It uns iuportant that thei

greaae be worked over and around the joint, assuring a smooth cn-tinuoue seal. _

It was noted during the first trial run that considerableleakage occurred through the soldered conectioa in the base plate.

Each hole that uns drilld and tapped proved to be a source ef

leake. In this investigatin hese leke uere stopped hy the

application of glyptal around the joints.

Rcconmendationss On the basis of the above discussion,the following recomendations are made.

·9&•

l. Thin coats of glyptal sealing cowound. should be used,

followed by a drying period of twenty·four hours.

Although seallng occurs when vacuun is isssediately

placed on wet glyptal, the seal does not hold for

following evacuations, small. air holes are produced

aa the drying occurs.

2. The base plate should be removed from the rest of the

apparatus and returned to Conas Cigarette Machine

Compaq tor resoldering. It is not advieed to attapt

the soldering unless considerable experience ie pos·

sessed, since the possibility of warping the hase plate

is great.

If the base place cannot be returned to Cmas

Cigarette Machine Conpany, smc fern of liquid soldercould be used. The joints should he sealed with a

permanent seal, not with sonething ae tenporsry as

glyptal.

3. Tygon tubing, reenforced, should be used for all

possible connections in preference to other types of

rubber tuhing.

Biffusion QggQgratiog: The operation of the diffusien pmup in this in-

vestigation was satisfactory. The pamp, as received from WarSurplus, contained oil that had previously been used and which

.95-

contained foreign matter. It was decided te use the oil in the punp

te determine the operating procedure of the pump rather than chance

ruining new eil. The foreign netter apparentlv had ne effect on the

operation of the pump.

The first time the punp was started, eonsideration degase·

ing time was necessary. After the first run it wae found that the

eil in the pump would degas as the rest of the eysten was degaseing

and ne additional time was necessary. In this investigation the

diftusion pump was turned en when the pressure indieated by the

preliminary vacmm gauge was less than 300 nierens. Twenty minutes

were needed to heat the eil te suffieient tuperature to allow the

diffusien pump te take over. The pressure would imediately drop

te less than 5 x 1.0% m. of Hg. when the pump did take ever.

It was found that a setting ef ninety velte en the power-

stat gave the maxinum pzmping speed. If the veltage was lower than

ninety er greater than ninety, the pressure within the still would

increase.goemendationnss The following reeousendations are node en

the basis of the above discussion.

l. The oil in the diffueion pup should be replaeed with

new eil. While the pump is enpty, the dirt and pieces

of glass that have eelleeted should be removed. _2. Fer the initial degassing ef the new oil, voltage to

the heaters should be applied slow]: to avoid vielent

.96-

bubbling of the oil. This should also be done if thepmp ls exposed to the atnosphere for long periods of

time without degassing.

§ell•jar Suggggg

During the runs made in this investigation, the bell•jar san‘

supported hy a length ef wire cord tiod to the frame and to the knobon the bell•jar. During the operation of the apparatus, it was foundthat the bell-jar did not need support, the pressure exerted free theoutside eas more than enough to hold it in place. With only the fore•

pup on it was found inpossible to move the bell•jar. However, failureof the puping system sold result in a loss of vaeunm an the jar

sould then fall.F t

1. It is reeonended that a more pernannt support, a

mechanical device, be provided for the bell·jar.

Figure 17, page 97, shows the details of a proposed

meehanical support. A metal cup, thin ealled, could

be made slightly larger than the knob on the bell- ‘

jar. A layer of rubber could be nolded to the in•

side of the cup. This could the be plaeed over the

knob.

A ssivel joint cold be made betsen the cup

and a threaded red. A cast iron braee, drilled and

threaded for the rod, cold be scresed on the rod.

A nnIIA

A

A -— Hiugod Joiuta AB —- Metal Cup

CDEPARIMIQJT OF CHPIMICAL ICNGINBLAING

PROPCJED B.€LL¤JAR .·$UR}‘(;»R'I‘.Jr«1w¤ by: ~· —” Date: 6-lf}-äöShecked by:§"g öcale: Nonex_yA_»;·¤ve<1 by:9w18 Figure No: 17

.93-

Two prongs of small diameter iron rod·cou1d be fastened

to the hraee with hinged connections, the prongs being

placed l80° apart- The ends of the prongs could be

. bent so that they would fit around the periphery of the

hase plate. They should be covered with rubher to avoid

damage to the base plate- A handle could then be put on

the other ends of the threaded rod- 9 „

In order to hold the bell•jar firmly in place, the

metal cup would be placed over the knob- The two pronga

wald then be hooked behind the hase plate- As the handle

was tuned, the brace would move awny from the cup, caus-

ing the prongs to tighte on the base plate ad the cup

to press on the knoh- By continued tightening of the

handle, any desired pressure could be ezerted on the be1l•

jar, holding it finnly against the base plate-

2- A wire screen, or cage, should be placed around the bell-

jar to deerease the danger to the operator in the eventA

the bell·jar is shattered- This cover could either be

separate, or could be inorporated with the proposed

bell·Jar support-V DistillationMean Free Path: The locatin of the bell•jar relative to

· the surface of the evaporator controlled the necessary vacuun neededt

to allow the oiledye feed nixture to dietill and be condensed within

. •99··

its mean free path. The bel1··jar was located two and one half incheeabove the surface of the evaporator. Aesusing that the nolecular di-

ameter of the oil used as feed was 10 x l0°8 cm., the calculated pres-

sure needed to assure moleeular conditions was 1 x 10°" mn. of Hg.This gave a mean free path of 8.3l• cm., calculated by the Maxwell

equetion.......3;.........L 2 J-E ud2n_

where

. L = Mean freepethincm.

N 4* Number of noleculee per unit volume, taken as

267 x 1019 moleculee per cc.d = Diueter of the moleeule 4

The distance from the bel1•jar to the evaporators was two and one

half inches, or 6.35 m. Distillation took place then within the _

mean free path of the oil noleeules, since pressure: less than S x10% ux. of Hg. were obteined.

4

Tgeraturex The tupereture maintained within the still' during the progress of the runs was 78°C. The boiling point of the

S.A.E. 30 motor oil was l.10'C., or 210*0., at atmoepheric pressure.

Since the oil distilled at a temperature of 78*0., a lowering of the

distillation temperature of 132*0. took place. This compared fa-

vorably with published data relative to lowering of distillation tem-

peratures.4

-199-

B. Limitatione

This investigation was eonducted under the following limltingT

conditions.

1. Pressurg: The operating pressure within the still was

naintained at less than 5 x 10-** m. of Hg. as indlested

by the prelininary, or hot wire vacuue geuge.

2. Temgggatgäz The temperature maintained within the still

was 78°C. as indicated by the Leeds and Horthrup tempera-

ture measuring potenticneter. The temperature read was

the temperature indicated by the thernocouple wire, the

hot junction of which was placed one inch below the ree·

idue outlet in the gutter. Thue, the tuperature ef the

residue leaving the roter was measured.

3. Mean Prä Path: The pressure maintained within the still

s corresponded to a nean free path of the oil noleculee of

8.3I. cm. (lt was assueed that the diaaeter of an oil

nolecule was 10 x 104 cm. This is probably high.) The

distance from the surface of the evaporater to the sur-

face of the condenser was two and one half inches,

6.35 cm.

L. Peä Ratg: The feed rate was eontrolled visuelly. The

rate which gave an extremely thin film of oil on the

evaporator was used. This variad with the BPM of the

evaporator.

5. §zgg: Ho analyees were eendueted to determine the

degree of separation obtained„ Visual neens were used

to determine that dietillation had occurred, the inside

ot the bel1·jar was seen to becone r•d•

•102·

V. CONCLUSIONS

Q1 the basis of the results obtained from the operation of

the five inch centrihsgal moleeular still designed and constructed

in this investigation, the following conclueions are drama

1. The punping equipuent installed will evacuate the volune

of the still, 10,300 ec. ( 10%), in tuenty ninutes, to epressure less than S x

lO°l‘m. of Hg.

2. The still allowed distillation of the feed mixture ¢<II·rposed of two quarts of motor oil, S.A.E. 30, and O.¤l0l•

graue of celanthrene red dye, at a temperature of 78'C.

and a pressure lese than 5 x 10% mu. of Hg.

3. The still lowered the distillation temperature of the

eil l32‘C., Iron 2l0°C. to 78°C.

-193-

V1. SUMLARI

The principle envolved in the molecular distillation process

is that of decreasing the pressure exerted on a suhstance thus lever-

ing the temperature at which it will distill• Boiling points of sone

materials can be lowered as much as l50°C„ The equipment operated in

this investigation, The Design and Construction of s Five Ich Centri-

fugel Molecular Still, sllowed such a decrease in distillation ten-

s perature.

The equipeent assemhled for use in this investigation con-

sisted of a still head, composed of a five inch roter, roter plate

and shaft, base plate, heater, collecting gutter, beering, and pack- ’

ing nat; vacuum pupe; notors; and niscellaneous glassuare necessary

to assmhle the apparatus•

In operation, the feed entered at the vertex of s rotating

evaporator heated hy the heater„ It passed over the surface of the

evaporato, a portion distilling, the rest roturning to a residue

centainer• The evaporated nolecules were cndensed and collected es

distillate• Separstion uns determined hy visual means, a red de,

celanthrene red, eas seen to separate from the nein portion of the

oil-dye feed ¤ixture•

Pressures obtained in the apparstus were less then5m¤•

of Hg. At these pressures and at a temperature of 78°C• die-

tillatien of the feed nixture took place. This uns a louering of

the distillation temerature of lß2'C•

-1;);,-

vll. BIBLIOGRAPH!

1. Bancroft, F‘.E. and Associated Electrical Industries, Ltd.

lmprovements in Vacuum Diatillation Apparatue. BritishPatent 1.28,719, May 17, 1935.

62. Baxter, J. G., Gray, E. L., and Tischer, A. 0. Molecular

Distillation. Preparation and Characterietics of Syn-thetic Conetant··Iield Mixturea. Ind. Eng. Chm. Q2,lll-2 (1937).

3. Bronsted, J. N., and Hevesy, G. von. On the Separation of theIsotopes of Mercury. Phil. Mag. Q, 31 (1922).

1.. . The Separation of theIsotopes of Mercury. Nature 106, lu; (1920).

5. Burch, C. R. and Van Dijck, W. J. D. The Theory and Develqnentof High Vacuun Distillation. J. Soc. Ghesa. Ind. j§, 39(1929).

6. and Bancrott, F. E. Iuproveuente in or Relating tovaeuua xmziiiazlaa. uam-. ggg, 729 (1928).

7. . Some Experimente on Vacuum Distlllation. Proc.Roy. Soc. (London), gg}, 271 (1929).

8. , Bancroft, F. E. , and Metropolitan-Vickers ElectricalCompany, Ltd. Inprevments in or Relating to Vacuum Die•

tillation. British Patent 303,078, Dec. 21, 1928.9. Burrows, G. J. The General Technique of Molecular üistillation II.

General Design of Molecular Distillation Equipment. J. Soc.

-105-1

can. ma. gg, so (1939).10. Caldvell, K. S. and Hurtley, N. H. The Dietillation ot Butter

\Fat, Cocoanut Oil, and Their Fatty Acids. J. Chem. Soc.

9 2,1. 853 (19G9)•11. Carothers, vi. H. and Hill, J. E11. Studies on rolymerization

and Ring Formation. J. Am. Chem. Soc. jg, 5279 (1930)•

12. Detuuiler, S. B., Jr. Abstracts of Ä1°t..1Cl88 and Patents on

Molecular or Short-Path Dietillation. U. S. Reg. Soybean

Ind. Products Lab., Urbana, lll., Dec. 191.1.

13. and Harkley, K. S. Laboratory Type Molecular

or Short Path Still for Vegetable and Animal Fats and

Gila. Ind. Eng. Chem. Anal. Ed. Q, 31.8 (191.0).

1.1.. übree, N. D. Molecular Distillation Theory of Elindnation

Curve. 1:.:1. chen. 33, 975 (1937).15. Favcett, E. W. H. The General Technique of Molecular Die-

tillation I. The Characteristics and Scope et the

Process. J. Soc. Chun. Ind. E, 1.3 (1939).

16. Hiciman, K. C. B. High Vacuum Short-Path Distillation - A

Review. Cha. Rev. 3;, 1(191.1.).J

17. . A Still tor Liquids ot High Boiling Point.

J. opt. see. Aa. gg, 69 (1929).18. — . Molecular Distillation Apparatus and Methods.

Ind. Eng. Chem. Q, 968 (1937).

-196-

19. Ibid. p. 1107.20. Hicknen, K. C. D. (To Eastman Kodak Oo.) Vacau Extractio

of Cod-Liver 011. U. S. Patent 1,925,559, Sept. 5, 1933.

21. Distillation Appa-

ratua. U. S. Patent l,9A2,858, Jan. 9, 193h.

22. . (To Distillation Products, Inc.) Vacuum

Distillation. U. S. Patent 2,2k9,526, July 15, 19hl. P

23. Vacuum

Distillation. U. S. Patent 2,210,928, Aug. 13, 1960..

2A. . Distillation Process. J. Franklin Inet.

Q2}, 215 (1936).25. . Degaasing Apparatus. U. S. Patent 2,197,539,

April 16, 19AO.

26. . New Advancaments in Ceutrifugal Steele.

mx. mag. chem. gg, 16 (191.7).

27. Jeans, J. “An Introduction to the Kinetic Theory of Gaeee“.‘ pp. 131. The Macmillan Co., N. Y. (l9&Ü).

28. Jewell, W., Head, T. H., and Phipp, J. W. The Application of

Molecular Distillation to the Concentration ot Vitamine.

J. Soc. Chen. Ind. jg, 56 (1939).

29. Knudsen, H. “The Kinetic Theory of Gasea“. pp. 21-26

Hethuen's Konographs o Physical Subjects. Messrs.

Hcthuen and Co., Landen 193h.

.197-

30. Lanmuir, I. Production of High Vacuum. J. Franklin Inst.

__l_Q_, 719 (19]-6)•

31. . Vacum Calcnlations. Phy. Rev. g, 329 (1913).

32. Quackenbush, F. H. and Steenbock, H. Effect of Time on Elimina-

tion Curvea. Ind. mg. Cha. Anal. Ed. 1}, 1.68 (191.3).

33. Washburn, E. w., Brunn, J. H., and Hicks, M. H. Apparatus and

Methods for the Separation, Identification and,Determina··

tion of Chemical Constituents of Petroleum. J. Res. Nat.

Bur. Standards g, 1.67 (1929).

31.. Ibid. p. 1.76.

35. Washburn, E. W. Chemical Constituents of Petroleum. Ind. äxg.

Chem. gj, 891—1. (1933).

36. , Brunn, J. H. and Hicks, H. M. Apparatus and

Methode for the Separation, Identification, and Detemina·

tion of Chemical Constituents of Petreleum. Research Paper

Aug. 1., 1928. Presented before Am. Cha. Soc., Sept. 1928;

Caldvell, K. S. and Hurtley, H. H. The Distillation of

Butter Fat, Cocoanut 011, and Their Patty Acid:. J. Chem.

Soc. Q, 853 (1909).

37. Wataman, H. 1. and Byholt, J. A. Destillatie Van Cocoaalie Bij

Zeer Lagen Drnk. Chem. Weekblad gg, 268 (1927)§ Caldwell,

K. S. and Hurtley, W. H. The Dizstillation of Butter Pat,

Cocoanut Oil, and Their Patty Acide. J. Cha. Soc. Q,853 (1909).

..1gg.

38. Ibid. p. 271..

39. Waterman, H. 1. and Oeaterhet, D. Polyneriutien ef Li¤•••d

Oil. Rec. trav. chem. 2, 895 (1933) (E¤&li•h); ¢¤ld¤ll»li. S. and Hurtley, ia. H. The Distillation of Butter Fat,

Cocoanut Oil, and Their Patty Acids. J . Chem. Soc. 2},

853 mw}.1.0. Wilson, R. E. Vacuum Distillation ot Petrolem. U. S. Patent

1.905,201, April 25. l933•

I

..199-

VIII. ACKNC:6L2;1JG¤;^sENTS

The author wishes to take this opportunity to express hisdeep appreciation to Professor Fred W. Bull, Associate Professor,for his advice and guidance during the course of this investigation.The help and advice of Dr. F. C. Vilbrandt, head of the ChemicalEngineering Department, and hr. JG. J. Geli, candidate for the degreeof Doctor of Philosophy, is also apprecieted.